Back to Notes

Wiki Operation Log

Wiki Operation Log

Append-only. Each entry format: ## [YYYY-MM-DD] action | description

Parse last 5 entries: grep "^## \[" log.md | tail -5

Actions: init | ingest | query | lint | synthesis | update

[2026-04-16] gap-fill | 16 missing pages created across 6 domains

Go (3 new): Context, HTTP Server, Testing DSA (3 new): Merge Intervals, Prefix Sum, Bit Manipulation System Design (3 new): Caching & Redis, Message Queues & Kafka, Consistent Hashing Python (3 new): Asyncio, FastAPI, Pydantic AWS (3 new): RDS, SQS & SNS, CloudWatch FrontEnd (1 new): TypeScript Updated: Go Topics, Python Topics, AWS Topics, FrontEnd Topics, index.md (124 pages)

Remaining gaps still open:

  • Go: sync.Pool/atomic, profiling/pprof, stdlib cheatsheet
  • DSA: Cyclic Sort, Fast & Slow Pointers dedicated page
  • System Design: YouTube/Netflix, Twitter Feed, WhatsApp, Google Docs, Hotstar, Maps/LBS
  • Python: pytest, dataclasses, context managers, type hints
  • AWS: DynamoDB, ElastiCache, API Gateway (AWS), ECS/EKS
  • AI/ML: LLM Fundamentals, Vector DBs, LangGraph

[2026-04-17] update | Weekly tracker created, old plan replaced, backlinks updated

  • Replaced Job Switch Plan 2026 (old 5-track plan) with revised single-focus plan
  • Created Weekly Progress Tracker (W06–W26) with milestones per week
  • Backlinks updated: DSA Topics, FrontEnd Topics, System Design Basics
  • Index updated: 133 pages, page_count corrected

[2026-04-17] gap-fill | 9 missing pages added (index: 124 → 133)

New pages:

  • DSA/SQL for Interviews — JOINs, window fns, CTEs, indexes, ACID, N+1, practice track
  • System Design/API Design Principles — REST, idempotency, pagination, versioning, status codes
  • System Design/Low Level Design — SOLID, 5 machine coding problems (LRU, rate limiter, parking lot, Splitwise, event emitter)
  • System Design/Docker and Kubernetes — conceptual reference, Dockerfile, K8s objects
  • System Design/Problem Designs/YouTube — transcoding pipeline, CDN, adaptive bitrate
  • System Design/Problem Designs/Twitter Feed — fan-out strategies, Snowflake IDs, Cassandra
  • System Design/Problem Designs/WhatsApp — WebSocket routing, message ordering, E2E encryption
  • System Design/Problem Designs/Google Docs — OT vs CRDT, collaborative editing deep dive
  • FrontEnd/Frontend System Design — typeahead, infinite scroll, real-time dashboard, state mgmt

[2026-04-17] update | DSA + SD progress reset to accurate state

DSA: Arrays & Hashing ✅, Two Pointers ✅, Sliding Window ✅, Binary Search ✅ — rest ⬜ SD: All 10 problems reset to ⬜ (interview-level bar: whiteboard in 45 min cold, not just read notes) Updated: Job Switch Hub, Interview Prep Hub

[2026-04-27] update | Vault-wide structural fixes (8 issues)

  • CLAUDE.md: SQL/ added to both directory tables, AWS description updated (MLA-C01 removed)
  • SQL/SQL Topics.md: index page created with practice tracker + concept checklist
  • CLAUDE.original.md: deleted (compress leftover)
  • Tech Stack Overview: AWS cert references removed, Bedrock production use noted
  • 2026 Goals: MLA-C01 → deferred, vectorbuilds.dev → deferred, learning section updated to new plan
  • Interview Prep Hub: DSA/SQL/LLD statuses updated to current state
  • index.md: page_count updated to 118, last_updated 2026-04-27
  • SQL for Interviews: practice tracker + concept checklist added

[2026-04-26] update | Weekly tracker cleanup + W05–W13 created (restart after Apr 13–26 break)

  • W01–W04 individual files deleted, consolidated into W01-W04 Archive (Mar 18–Apr 12).md
  • Created: W05 (Apr 27–May 3) with full daily plan, W06–W13 with track plans + metrics
  • Updated: Weekly Progress Tracker (phases rebased to W05 start Apr 27), Job Switch Hub (current state)
  • W05 focus: Linked Lists + URL Shortener (interview-ready) + SQL start
  • Apply sprint unchanged: first applications Jun 1 (W10)

[2026-04-17] update | Prep plan + synthesis pages overhauled post career-clarity session

Updated: synthesis/Job Switch Hub.md, synthesis/Interview Prep Hub.md

Key changes:

  • Honest profile assessment added (what he actually built vs claimed)
  • Prep priority revised: 7 core topics + 5 for full-stack (not 15 parallel tracks)
  • Apply date moved up: Jun 1 not Jul 18 (AI startups first)
  • Job links added: Sarvam AI (4 live postings), Razorpay, PhonePe, CRED, BrowserStack, Zepto, Swiggy, Juspay
  • SQL section added to Interview Prep Hub (fintech companies test this — was missing)
  • LLD/Machine Coding moved up to W06 (was W10-W11 — too late)
  • FastAPI internals section added with code patterns
  • API Design principles section added
  • AskTGE designated as primary interview story everywhere
  • Backend project added as explicit milestone (closes architecture credibility gap)
  • Dropped: AWS MLA-C01 cert, Go mastery track, blog — all post-offer

[2026-04-17] update | Resume rewritten — accuracy pass + AskTGE added

Saved to: Private/Resume/Resume_2026.tex

Key changes from original:

  • Headline: "full-stack + distributed systems" → "backend systems + real-time + AI integrations"
  • AskTGE LLM optimization added as first HCLTech bullet (was missing entirely)
    • DynamoDB schema redesign + API latency 2s → 200ms + hallucination fixes + summarization
  • WebSocket bullet reframed: "Engineered geospatial engine" → "Built driver tracking FE + FastAPI initial load"
  • ML Thresholding: "Architected and deployed" → "Implemented parallel processing + per-KPI failure isolation"
  • Skills cleanup: removed Vue.js, D3.js, UI/UX Design, C++, Deep Learning, Pandas, Node.js
  • Skills added: DynamoDB, WebSockets, LangChain, RAG Pipelines, PostgreSQL

[2026-04-16] init | Karpathy wiki architecture implemented

Bootstrapped wiki infrastructure:

  • Added three-layer architecture (raw sources / wiki / schema)
  • Created raw/ for immutable source documents
  • Created synthesis/ for cross-domain LLM-maintained pages
  • Created index.md — full wiki catalog (108 pages)
  • Created log.md — this file
  • Created Uncategorized/ directory (was referenced but missing)
  • Updated CLAUDE.md with ingest/query/lint workflows
  • Created 5 synthesis pages:
    • synthesis/Job Switch Hub.md — 50+ LPA goal tracking
    • synthesis/Interview Prep Hub.md — aggregated interview prep
    • synthesis/Tech Stack Overview.md — positioning guide
    • synthesis/Concurrency Deep Dive.md — Go + Python + distributed
    • synthesis/LLM & AI Stack.md — AI company targeting

[2026-04-27] update | SQL resource sync + Binary Search W04 status

  • synthesis/Interview Prep Hub.md — SQL Practice section: HackerRank → DataLemur + LeetCode SQL 50 (matches SQL for Interviews.md)
  • Private/Project_2026/Weekly Tracker/Week 05.md — all HackerRank SQL refs → DataLemur Easy + LeetCode SQL 50
  • DSA/DSA Techniques for Interviews/Binary Search.md — marked LC 704, 74, 153, 33, 981, 875 as ✅ W04; LC 4 (Median of Two Sorted Arrays) as ⬜ Deferred

[2026-04-27] ingest | The Day a Google L7 Engineer Tore My System Design to Shreds

Source: Medium / Cloud With Azeem — L7 system design interview lessons via URL shortener war story

Pages created:

  • System Design/Senior SD Mindset.md — Pattern Trap, Hot Key problem, cache hit rate formula, failure-first thinking, trade-offs, prep methodology

Pages updated:

  • System Design/Problem Designs/Design a URL shortener.md — added cache hit rate math, hot key problem + fixes, 3 new interviewer checklist items, Related link
  • synthesis/Interview Prep Hub.md — added L7 signals block (cost every component, hot key, cache math, failure-first, "we don't need X yet")

raw/ marked: ingested: true, wiki_pages_updated set index.md: +1 page (119 total), new SD entry, raw sources table populated

[2026-04-28] ingest | 3 articles processed — Redis, AI Concepts, Real SD Interviews

Redis Architecture Deep Dive (Nitesh Thakur)

Pages updated:

  • System Design/Caching & Redis.md — added "Redis Architecture Deep Dive" section: single-threaded model + I/O multiplexing, data structure internals (string encodings, list O(N) gotcha, sorted set dual-storage, hash ziplist 3x savings), persistence deep dive (RDB fork latency, AOF growth + BGREWRITEAOF, production guide), Lua scripts for atomicity, pipelining (100x), transactions no-rollback gotcha, memory management (eviction policies, fragmentation), monitoring (SLOWLOG, INFO), cluster hash tags

20 Most Important AI Concepts (Deep concept)

Pages created:

  • AI & ML/LLM Fundamentals.md — transformer stack (tokenization → embeddings → attention → transformer), LLM fundamentals (context window, temperature, hallucination), training optimization (fine-tuning, RLHF, LoRA, quantization), prompting (prompt engineering, CoT), AI systems (RAG, vector DB, agents, diffusion) Pages updated:
  • AI & ML/AI & ML Topics.md — checked off 7 items from LLM fundamentals checklist, added LLM Fundamentals to nav

Real SD Interviews at 10+ Companies (Freeze Francis)

Pages created:

  • System Design/Real Interview SD Questions.md — 10 real questions (pagination, WhatsApp, cross-DC replication, SIM store, saved-search notifications, push notifications, load balancer, bike rental, subscription payment), deep dives on pagination + chat + payment, patterns observed (push systems, LLD gap, domain-specific, seniority bar, tooling matters) Pages updated:
  • synthesis/Interview Prep Hub.md — added real interview patterns block

All 3 raw files: ingested: true index.md: +2 pages (121 total), all new pages + raw sources indexed

[2026-05-03] update | SQL knowledgebase expanded — 8 topic files added

Pages created:

  • SQL/Basics.md — SELECT, WHERE, ORDER BY, DISTINCT, NULL, COALESCE, CASE WHEN, data types, execution order
  • SQL/JOINs.md — INNER/LEFT/RIGHT/FULL OUTER/CROSS/self-join, anti-join patterns, NOT IN NULL gotcha
  • SQL/Aggregates.md — COUNT/SUM/AVG, GROUP BY, HAVING vs WHERE, conditional aggregates, ROLLUP
  • SQL/Subqueries & CTEs.md — scalar/correlated/EXISTS subqueries, CTEs, recursive CTEs with execution model
  • SQL/Window Functions.md — ROW_NUMBER/RANK/DENSE_RANK/NTILE, LAG/LEAD, running totals, frames, FIRST/LAST_VALUE
  • SQL/Indexes & Performance.md — B-tree/composite/partial/covering indexes, EXPLAIN ANALYZE, query optimization
  • SQL/Transactions.md — ACID verbal, isolation levels, SELECT FOR UPDATE, deadlocks, MVCC, optimistic locking
  • SQL/String & Date Functions.md — string ops, LIKE/ILIKE, date math, DATE_TRUNC, COALESCE/NULLIF, CAST, JSON

Pages updated:

  • SQL/SQL Topics.md — Quick Nav updated with links to all 8 new files
  • index.md — SQL section updated (2 → 10 pages), page_count 121 → 129

[2026-05-03] analysis | Senior analyst portfolio review (May 2026) — refreshed trackers from Monthly Data dump, combined corpus ₹40.32L (+₹8.56L since Apr 2). P1 +13.66%, P2 +5.79%. Updated Investment Tracker, Portfolio 2 Tracker, Combined Summary. Added global investment mandate (US ETFs ITA/SMH, SGBs, Indian REITs, tax-free PSU bonds) for 3-5yr horizon.

[2026-05-10] update | Window Functions — added ROWS vs RANGE INTERVAL example (date-gap gotcha for rolling-7-day sums, INTERVAL N DAY PRECEDING vs N PRECEDING rows).

[2026-05-10] ingest | xoraus/CrackingTheSQLInterview (115 Q&A SQL interview guide) — covered: DDL/DML/DCL/TCL classification, TRUNCATE vs DELETE vs DROP, AUTO_INCREMENT, all 6 constraints (NOT NULL/UNIQUE/PK/FK/CHECK/DEFAULT) + ON DELETE CASCADE/SET NULL/RESTRICT, Views (incl. materialized), Stored Procedures, Cursors (anti-pattern), Triggers (BEFORE/AFTER, NEW/OLD), SQL Injection defense (parameterized queries, identifier injection), GRANT/REVOKE, roles, least privilege. 4 new wiki pages created, SQL Topics nav + concept checklist updated, page_count 132 → 136.

[2026-05-10] progress | Last 2 weeks (W05 Apr 27–May 3 + W06 May 4–10) — SQL completed: DataLemur "Learn SQL" course done (SELECT/JOINs/aggregates/subqueries/window fns); LeetCode SQL 50 fully done (53 problems — full list in SQL Topics). All classic interview patterns covered cold (2nd highest salary, dept top-N, consecutive nums, dept highest salary, delete dup emails). W05 + W06 retrospectives + concept checklists updated. SQL track: interview-ready bar reached. Carry to W07: Trees DSA, Rate Limiter SD, Resume polish (deferred during SQL push).

[2026-05-11] carry | W05+W06 unfinished items rolled into W07 — Trees DSA (14 LC), Linked Lists DSA (9 LC), URL Shortener + Rate Limiter SD (carry W05–W06), Resume polish + send (carry W05). SQL track marked done — maintenance only.

[2026-05-16] progress | W1 gate closed — 15/15, PASS STRONG

LC 4 (Median of Two Sorted Arrays) solved in 30 min — senior bar. Binary search on smaller array, correct partition logic. Updated: Private/Project_2026/Weekly Tracker/Week 01 (May 11–17) — Validation Gate.md — #15 marked Done, Score: 15/15 PASS Updated: synthesis/Coach Hub.md — W1 status summary reflects 15/15 clean pass W2 direction: Linked List, 11 NeetCode problems, May 18–24.

[2026-05-16] update | Coach rearchitected — reads vault files directly, no stale memory

Problem solved: Memory files were stale (showed 0/15 problems solved; vault showed 14/15 with detailed notes).

New file created:

  • synthesis/Coach Hub.md — LLM-maintained coach state hub. Holds current_week_file pointer (updated each Sunday retro), W1 status summary, applications table, mocks done. Coach reads this first, then follows the pointer to the actual weekly tracker file.

coach.md rewritten (~/.claude/commands/coach.md):

  • Reads from vault: synthesis/Coach Hub.md → Private/Project_2026/Weekly Tracker/<current week>.md → synthesis/Job Switch Hub.md → Private/Project_2026/Projects Done.md
  • Keeps only feedback_dsa_solution_review.md in memory (Claude behavior instruction, not personal data)
  • Writes back only to synthesis/Coach Hub.md after retro/scorecard (never to Private/)

Hook fixed (.claude/settings.json):

  • Added vault-path guard — hook now only fires on files inside the vault directory
  • Previously over-triggered on ~/.claude/ files (coach.md, memory files)
  • Added /.claude/ and /memory/ to skip list

index.md: +1 page (Coach Hub), page_count 125→126

[2026-05-16] update | DSA Techniques overhaul — numbered files, index fixed, 3 notes upgraded

index.md:

  • Added 3 missing DSA Technique entries: Arrays & Hashing, Greedy, Math & Geometry
  • Fixed section header counts: DSA (15→27), Go (20→23), Python (7→10), AI & ML (3→4), FrontEnd (9→11)
  • page_count 122→125, last_updated 2026-05-16

DSA/DSA Techniques for Interviews/ — all 18 files renamed with numeric prefixes (01–18): 01. Arrays & Hashing → 18. Trapping Rain Water Visual (NeetCode curriculum order)

Cross-links fixed: all [[links]] inside technique files updated to numbered names

Technique notes upgraded:

  • 05. Binary Search.md — added rotated array pattern code (Pattern 4), find-min pattern (Pattern 5), bisect module examples, interview opener, while left<right vs <= explanation; removed placeholder Notes section
  • 11. Graphs.md — added grid DFS/BFS template (Pattern 6, LC 200), directed cycle detection with in-stack set (Pattern 7); removed placeholder Notes section
  • 08. Tries.md — added wildcard/dot search (Pattern 2, LC 211), Word Search II full code with trie pruning (Pattern 3, LC 212), expanded Senior Tricks

Verified excellent (no changes): Arrays & Hashing, Two Pointers, Sliding Window, Stacks, Linked Lists, Trees, Heap, Backtracking, Dynamic Programming, Greedy, Merge Intervals, Math & Geometry, Bit Manipulation, Prefix Sum

[2026-05-19] progress | SD Theory: Introduction to System Design completed

Updated: SD Theory Curriculum (Introduction ✅, completed count 1→2/12), Job Switch Hub (Introduction row ✅), Coach Hub W02 (SD progress 1/3 done). Remaining W2 SD: Caching + Asynchronous Systems.

[2026-05-19] update | W2 weekly tracker created

Created: Private/Project_2026/Weekly Tracker/Week 02 (May 18–24) — Linked List.md — 15 LL problems table, SD theory tracker (4 sections), BA drills tracker, drill Qs tracker, day-by-day plan (7h Sat+Sun), decision gate, hours log, retro section.

[2026-05-19] update | Vault rules updated + hook removed + W1 retro written

CLAUDE.md: LLM now owns Private/Project_2026/Weekly Tracker/ — can write retro data, problem scores, gate verdicts. Blanket Private/ ban updated to specific exclusions (Personal/, Archive/, Interview Questions/). Enforcement rule changed from PostToolUse hook to self-enforced CLAUDE.md rule. .claude/settings.json: vault_log_enforcer PostToolUse hook removed. Private/Project_2026/Weekly Tracker/Week 01 — Validation Gate.md: retro section filled — 15/15, ~192 min, weakest Binary Search, what slipped (behavioral), what surprised (speed), W2 decision (LL 15 problems), fix (behavioral Day 1).

[2026-05-19] update | W2 daily schedule added — 7h/day weekend load

Coach Hub W02: day-by-day plan added. Weekdays Tue–Fri ~2.5–3h each (light DSA + SD theory + behavioral). Sat–Sun 7h each (heavy DSA hards + Resiliency + Essentials + retro). Total W2: ~26.5h.

[2026-05-19] update | W2 plan finalized — 15 LL problems, 4 SD sections, BA + drill questions

W2 scope: Linked List expanded 11→15 (added LC 876, 92, 234, 328 — covers slow/fast pointer, sublist reversal, palindrome, odd-even patterns). SD theory 4 sections this week (Caching + Async + Resiliency + Essentials, ~9h). BA drills and pre-screen drill questions formalized as explicit W2 items. Introduction to System Design moved to W1 completion (not W2). W1 SD total: Databases + Introduction. SD curriculum schedule rewritten: W1=2, W2=4, W3=3, W4=3 sections — all 12 done by W4 end, W5 buffer, W6 problems start on time. W2 decision gate updated for 15 problems: 13–15 pass strong, 10–12 patchy, <10 fail. Updated: SD Theory Curriculum, Coach Hub W01+W02, Job Switch Hub.

[2026-05-19] update | SD theory W2 increased — 3 sections/week pace locked in

SD Theory Curriculum completion targets rewritten: W2 = Intro + Caching + Async (~5h), W3 = Resiliency + Essentials + Networking, W4 = Consumer + Observability + Security, W5 = Utilities + Critical + buffer. All 12 sections done by W5 end, W6 SD problems on schedule. Coach Hub W02: SD theory target updated to 3 sections this week.

[2026-05-19] update | SD theory curriculum expanded + DSA theory block rule added

SD Theory Curriculum: 9 → 12 sections. Added Networking & Protocols, Observability, Security Basics. Added API Design Principles cross-link. Completion target table updated (new sections target W4–W5). Job Switch Hub: standing rule added — every new pattern week starts with 1–2h theory block (vault technique page + NeetCode intro) before problems. Coach Hub W02: pattern theory block added (1.5h: Linked List vault page + NeetCode intro, covers dummy node / slow-fast pointer / in-place reversal before touching LC 206).

[2026-05-19] update | W1 retro closed — SD theory databases done, vault state finalized

SD theory Databases section completed (relational, isolation levels, scaling, sharding, non-relational, DB selection). Created: System Design/SD Theory Curriculum.md — 9-section theory tracker, Databases ✅, rest ⬜. Target: full theory by W6 (Jun 21). Updated: synthesis/Coach Hub.md — W01 SD theory marked ✅, W02 carry-over debt reduced to STAR drills + pre-screen Qs only. Updated: synthesis/Job Switch Hub.md — W1 DSA gate checkbox ✅ (15/15 PASS STRONG), SD Theory Foundation section added with curriculum summary table. Updated: index.md — SD Theory Curriculum added, System Design 21→22 pages, page_count 126→127.

W1 retro summary: DSA 15/15 STRONG, SD theory databases done, STAR drills + pre-screen Qs carry to W2. Mocks: 0 (expected). Apps: 0 (expected — W3 target).

[2026-05-19] update | W1 scorecard closed, Coach Hub advanced to W2

W1 final score: 15/15 on time, solo (~192 min / 420 min target). Gate: PASS STRONG. Weakest: Binary Search (Search Rotated 28/30). Strongest: Two Pointers + Stack. Non-DSA carried to W2: STAR drills, pre-screen Q drafts, SD theory basics. Updated: synthesis/Coach Hub.md — W01 status finalized, current_week_file → W02 (May 18–24), Past Week Summaries table updated, W02 status + problem set + decision gate added. W2 direction: Linked List, 11 NeetCode problems, target Sun May 24 gate.

[2026-05-12] lint | Full vault audit + reorganization — 9 stale files removed, 6 files updated

Removed:

  • AWS/ML Associate Exam/ (7 files) → archived to Private/Archive/Deferred/AWS-ML-Cert/ (kill list: post-offer)
  • System Design/Backend 101/Concurrency/Concurrency.md — stub (20 lines), superseded by synthesis/Concurrency Deep Dive.md
  • System Design/Backend 101/Data Base/Basics.md — stub (19 lines), superseded by System Design/System Design Basics.md + SQL/

Updated:

  • index.md — removed dead links (PROJECT_2026_CONTEXT, Job Switch Plan 2026), updated synthesis page dates (→ 2026-05-12), fixed AWS section (ML exam removed), added Private Trackers with AskTGE + Coach file refs, page_count → 122
  • Private/Project_2026/Quick Links.md — moved killed resources (AWS cert, LangChain, Frontend SD) to new Deferred section; active list: NeetCode, Boot.dev, Gaurav Sen, Layrs.me
  • Private/Project_2026/Projects Done.md — populated with all 4 war stories: AskTGE perf audit, Chat History DynamoDB, TGE WebSocket, Splunk ML Thresholding
  • Private/Interview Questions for Project or behaviour/Past Experience Q&A For Interview.md — added YAML frontmatter + nav table
  • DSA/DSA Topics.md — added Greedy Algorithms (§10), Intervals (§11), Math & Geometry (§12) pattern tables; added NeetCode 150 progress tracker at bottom

[2026-05-21] update | Coach Hub W02 solve count updated to 5/15 (LC 206, 21, 141, 876, 143 done)

[2026-05-21] update | W02 tracker: LC 19 solved (20:00 / 25 min target). Coach Hub updated to 6/15.

[2026-05-21] update | W02 tracker: LC 138 solved (tail bug noted). Coach Hub updated to 7/15.

[2026-05-22] update | W02 tracker: LC 2 solved (naming note: quo→digit). Coach Hub updated to 8/15.

[2026-05-22] update | W02 tracker: LC 92 solved (20:00 / 25 min target). Coach Hub updated to 9/15.

[2026-05-22] update | W02 tracker: LC 234 solved. Coach Hub updated to 10/15.

[2026-05-23] update | W02 tracker: LC 328 FAIL (34/25 min timer blown), LC 287 solved (24/35 min). Coach Hub updated to 11/15 (10 on time, 1 fail).

[2026-05-23] update | W02 tracker: LC 146 solved (36/40 min). Coach Hub updated to 12/15 (11 on time, 1 fail).

[2026-05-24] update | W02 tracker: LC 23 solved (15/45 min, asterisked — heapq taught mid-session). Coach Hub updated to 13/15.

[2026-05-24] update | W02 tracker: LC 25 solved (43/50 min). All 15 problems done. Final score 13/15 on time (1 fail 328, 1 asterisked 23). Coach Hub updated.

[2026-05-25] scorecard | W02 PASS STRONG (13/15 DSA on time). Coach Hub → W03 pointer set. Job Switch Hub: W2 ✅, Caching SD ✅. Carry to W3: Async+Resiliency+Essentials SD, BA drills 11 reps, Drill Qs rehearsal.

[2026-05-25] retro | W02 retro written to weekly tracker. What surprised: friend visit + heavy work load cut hours (not a W1 factor). Fix for W3: fixed time slots for non-DSA work.

[2026-05-25] update | W03 weekly tracker created — Trees pt 1, 13 problems, referral DMs added, BA drills reduced to 3 reps/week minimum.

[2026-05-27] update | W03 tracker: LC 226 solved (3/10 min). Coach Hub updated to 1/13.

[2026-05-30] update | W03 tracker: LC 235 (7 min), LC 102 (5 min), LC 199 (5 min) solved. Running total 9/13.

[2026-05-30] update | W03 tracker: LC 1448 (7 min, self-corrected), LC 98 (4 min), LC 105 (15 min) solved. Total 12/13 on time.

[2026-05-31] update | W03 tracker: LC 230 (6 min) solved. Final score 13/13. Score line updated.

[2026-05-31] retro | W03 PASS STRONG (13/13, ~86 min / 265 min). Coach Hub → W04 pointer set. Job Switch Hub W3 ✅. W04 tracker created. Carry: STAR reps + referral DMs (mandatory Jun 7).

[2026-06-03] update | W04 tracker: LC 114 note updated — solution created new TreeNode objects (in-place violation). Marked Yes* with fix note: Morris O(1) or stack-relink.

[2026-06-03] update | W04 tracker: LC 297 note updated — BFS correct, spurious inner for-loop in serialize. DFS + iter()/next() is cleaner canonical.

[2026-06-04] update | W04 tracker: LC 103 solved. Style: cnt→boolean flag, list(reversed())→slicing. Logic at senior bar.

[2026-06-06] update | W04 tracker: LC 113 solved (7:30). Minor: subtract-then-check-zero cleaner than leaf-val==target. Also solved LC 112 warmup (4:30).

[2026-06-06] update | W04 tracker: LC 116 solved (15:00). DFS correct, O(1) iterative is canonical. LC 117 follow-up noted.

[2026-06-06] update | W04 tracker: LC 117 (arbitrary tree follow-up) solved as bonus. O(1) scan-forward approach, senior bar.

[2026-06-06] update | W04 tracker: LC 450 FAIL (36 min vs 30 min). Hybrid iterative/recursive approach. Score 6/7. Gate: PASS STRONG → W5 Tries + Heap pt 1.

[2026-06-07] update | W04 tracker: LC 2196 (bonus) added — 20 min, setdefault + set difference pattern noted.

[2026-06-13] update | DDIA Ch1 "Trade-Offs in Data Systems Architecture": formatted notes (frontmatter, headings, fixed broken OLTP/OLAP table); converted 2 PNG figures (ETL warehouse flow, cloud-vs-self-host spectrum) to Mermaid.

[2026-06-14] create | synthesis/DDIA Interview Map: chapter→SD-topic map, reading protocol, design pairings + hands-on experiments. Linked into Job Switch Hub. Added "Interview Trade-Offs (Ch1)" section to DDIA Ch1 note + formatted appended Cloud Native section/Table 1-2. Indexed DDIA Ch1 note + new synthesis page (page_count 127→129).

[2026-06-14] update | Corrected synthesis/DDIA Interview Map to DDIA 2nd-ed (2025) 14-chapter layout (prev used 1st-ed numbering). Added ch2 Nonfunctional Requirements, split Sharding/Distributed/Consensus, added highest-ROI chapter list. Updated Ch1 note with 2nd-ed section checklist (covered + remaining).

[2026-06-14] create | .claude/commands/ddia.md skill — post-chapter processor: format note, fact-check/rewrite weak concepts, add Interview Trade-Offs + next hands-on exercise, wire into DDIA Interview Map + index + log. Invoke with /ddia [chapter file|number|empty].

[2026-06-15] update | DDIA Ch1 (via /ddia): formatted newly-added back-half (Cloud Native internals, Distributed vs Single-Node, Microservices/Serverless) — •→GFM bullets, fixed run-ons + typos (cloud-native, idle); rewrote garbled storage/compute-separation + multitenancy paragraphs. Added 5 new Interview Trade-Offs (storage/compute split, distributed vs single-node, microservices vs monolith, serverless) + "Next: Hands-On Exercise" (dashboards design drill + Postgres OLTP-vs-aggregate experiment). Coverage now 12/15 sections; 3 short ones left.

[2026-06-15] update | /ddia skill: added incremental-run handling (Step 1.5) — git-diff to find new/raw spans, leave polished prose untouched, never fabricate unwritten sections, append-don't-replace for trade-offs. Report now states full vs incremental + sections actually touched.

[2026-06-19] update | DDIA Ch1 (via /ddia, incremental): formatted newly-added "Data Systems, Law & Society" (run-on prose→bold-term bullets across 4 subsections; fixed typo "data minimization. data minimization", missing periods, double spaces) + reworked raw "Cloud Computing vs Supercomputing" bullets into a GFM comparison table (workload/fault-model/comms/topology/geography). Added 4 Interview Trade-Offs (cloud vs HPC fault model, data minimization as liability, immutable logs vs right-to-erasure) + 1 more. Coverage now 14/15 sections — only book Summary skipped (complete). Updated Interview Map chapter-notes bullet + index.md entry to (complete).

[2026-06-20] update | DDIA Ch2 "Defining Nonfunctional Requirements" (via /ddia, full first pass): added frontmatter + coverage checklist; structured whole file into headings + bold-term bullets; fixed typos (referes, resoruces, ans, "file seconds", SQL aliases, "99.0th"→p999) and run-on prose. Converted 4 ![[image]] embeds → Mermaid: users/posts/follows erDiagram, response-time-vs-throughput xychart (hockey stick), request-lifecycle latency-breakdown flowchart, and a percentile-ladder table+callout. Added Interview Trade-Offs (Ch2): NFRs-first, fan-out write-vs-read, celebrity hybrid, lossy sampling, percentiles-not-mean, client-side measurement, throughput headroom, metastable-failure defenses. Added Hands-On (home-timeline design drill + p99-vs-throughput load-test). In progress — Reliability/Scalability/Maintainability not yet written. Indexed (page_count 129→130), updated Interview Map chapter notes. 4 source PNGs now orphaned (awaiting delete confirmation).

[2026-06-21] update | DDIA Ch2 "Defining Nonfunctional Requirements" (via /ddia, incremental): formatted newly-added Use of Response Time Metrics, Reliability & Fault Tolerance, Scalability, Maintainability — run-on prose→bold-term bullets/callouts; fixed typos ("scaling our"→out, invovle, "with in", approximatly, drivers→drives) + a glued "---" separator. Converted tail-latency-amplification image → Mermaid fan-out flowchart (slowest backend dominates); added SLO vs SLA distinction + percentile-estimation libs + "don't average percentiles" note. Reviewed/tightened fault-vs-failure, SPOF, chaos engineering, HW/SW/human faults, shared-memory/disk/nothing, scalability principles. Appended 6 Interview Trade-Offs (tail-latency amplification, fault vs failure, reliability spend, vertical vs horizontal, don't-over-engineer-scale). Chapter now complete (only book Summary skipped). Updated Interview Map + index to (complete). Image Pasted-image-20260620165312.png now orphaned (awaiting delete confirmation, along with 4 earlier ones).

[2026-06-21] update | DDIA Ch2 "Defining Nonfunctional Requirements" (via /ddia, incremental): formatted three newly-expanded Maintainability subsections — Operability (run-on→bold-term bullets + monitoring/no-machine-dependency/predictability list, fixed lowercase "more automation not equal to" fragment), Simplicity (essential vs accidental complexity, abstraction, design patterns/DDD), Evolvability (Agile/TDD/refactoring, coupling↔changeability). Kept prior 3-bullet overview as section intro. Appended 1 Interview Trade-Off (maintainability: kill accidental complexity). Chapter remains complete. No wiki status change (already marked complete earlier today).

[2026-06-21] update | DDIA Ch2 cleanup: deleted 5 orphaned source PNGs (all converted to Mermaid — schema erDiagram, throughput xychart, latency-breakdown flowchart, percentile table, tail-amplification fan-out) after confirming no remaining ![[ ]] embeds. Enriched home-timeline design drill with reliability follow-up (fan-out worker crash → SPOF, idempotent/no-dup takeover) + scalability follow-up (shared-nothing sharding by user_id, hot-shard for celebrities). Hands-on (p99 vs throughput) left unchanged — still the chapter's core experiment. xychart confirmed rendering.

[2026-06-21] update | Progress review + REPLAN. Found Jun 8–21 deviation (no DSA; all time on DDIA Ch1/Ch2 notes + /ddia skill — off critical path). Status: DSA strong thru Trees pt2 (W4, 6/7 PASS); apps/DMs/mocks/STAR all 0. Replanned per Vatsal into 8-week sprint (Jun 22–Aug 15): finish ~90 NeetCode→150 + ~8 HLD/SD problems + Redis playlist + remaining SD theory, THEN apply mid-Aug. Decisions: apps/DMs→August, DDIA folded into SD (1ch/2wk), STAR dropped from gate, DSA compress+slip 1wk. Writes: created Week 05 (Jun 22–28) Tries+Heap tracker (S1, +URL Shortener +Redis); rewrote project_dsa_sd_curriculum.md (S1–S8 table) + project_job_switch_2026.md phases + project_current_week.md memory; Coach Hub (W04 final summary + S1 pointer + replan banner); Job Switch Hub (W4✅, SD theory done, DDIA folded, phase map → 8-wk sprint, apps→Aug, URL Shortener→in progress).

[2026-06-21] update | Sprint detail-out: added DDIA per-chapter targets to curriculum (Ch6 Replication S1–S2, Ch7 Sharding S3–S4, Ch8 Transactions S5–S6, Ch9+10 S7–S8; Ch3/Ch4 stretch). Created "Sprint Plan (Jun 22–Aug 15)" master tracker file — 8-week task checklists (DSA per-pattern problem lists + SD problem + Redis topic + DDIA chapter per week) + sprint scoreboard. Coach skill invoked for sprint-readiness assessment.

[2026-06-21] update | Replaced single master sprint file with 7 separate weekly tracker files (per Vatsal request): Week 06 Backtracking (S2), Week 07 Graphs (S3), Week 08 Adv Graphs+Greedy (S4), Week 09 1-D DP (S5), Week 10 2-D DP (S6), Week 11 Intervals+Math/Geo (S7), Week 12 Bit Manip+Sprint Close (S8). Each: goals, full NeetCode problem table w/ LC#/diff/target, SD+Redis+DDIA tracker, compact daily plan, decision gate, hours log, retro stub. Week 12 includes sprint-complete retro + flip-to-apply checklist. Deleted "Sprint Plan (Jun 22–Aug 15)" master file.

[2026-06-21] update | S1 tracker: LC 208 Implement Trie solved (6:00 / 25 min). Senior bar. Polish noted: extract _find helper to DRY search/startsWith (reused 211/212). Running 1/10.

[2026-06-21] update | S1 tracker: LC 211 Add and Search Words solved (13:00 / 30 min). Senior bar — reused _find helper from 208, DFS branch on dot. Nit: any([..])→generator. Running 2/10.

[2026-06-21] update | S1 tracker: LC 212 Word Search II solved (time TBD). Senior bar — fully optimized: single Trie-in-lockstep DFS + leaf-prune del + visited # + is_end dedup. Tries bucket 3/3. Running 3/10.

[2026-06-23] update | S1 tracker: LC 703 Kth Largest in Stream solved (3:00 / 15 min). Senior bar, min-heap size k. heapq reflex confirmed (W2 flag closed). Running 4/10. (LC 212 time still pending.)

[2026-06-23] update | S1 tracker: LC 212 time filled = 24:00 / 45 min (hard, under target). Tries bucket 3/3 all senior bar.

[2026-06-23] update | S1 tracker: LC 1046 Last Stone Weight solved (7:30 / 15 min). Senior bar, negate-for-max-heap, correct y-x sign. Running 5/10.

[2026-06-23] update | S1 tracker: LC 973 K Closest Points solved (9:00 / 25 min). Senior bar, max-heap size k O(n log k). Nit: drop sqrt, compare squared dist. Running 6/10.

[2026-06-23] update | S1 tracker: LC 215 Kth Largest in Array solved (time TBD). Senior bar heap O(n+k log n). Flagged: write quickselect O(n) cold (shared 215/347/973). Running 7/10.

[2026-06-23] update | S1 tracker: LC 215 time filled = 5:00 / 25 min.

[2026-06-24] update | S1 tracker: LC 621 Task Scheduler FAIL (43:30 / 30 min). Simulation correct but over time — should use O(N) formula max(len, (max_freq-1)*(n+1)+num_max). First miss of week. Running 7/8 on time, 8/10 attempted.

[2026-06-24] update | S1 tracker: LC 355 Design Twitter FAIL* (36:00 / 35 min + hints on getNewsFeed). Design correct/optimal (lazy k-way merge) but merge pattern not reflex — same gap as LC 23 (W2). Action: drill 23+355. Ties to S3 Twitter Feed SD. Running 7/9 on time, 2 fails.

[2026-06-24] update | S1 tracker: LC 295 Find Median from Stream solved (time TBD), solo no hints. Senior bar two-heap canonical. DSA bucket complete 10/10 attempted (8 on time, 2 fail: 621 formula, 355 merge-pattern). Finished Wed Jun 24, target was Sun.

[2026-06-24] update | S1 tracker: LC 295 time = 9:00 / 40 min (hard, solo, no hints). DSA bucket fully scored.

[2026-06-24] update | S1 tracker: scheduled Fri Jun 26 as merge-pattern review day — LC 23→355 back-to-back solo + 621 formula cold rewrite. Closes the recurring k-way-merge gap (W2 LC23 + S1 LC355).

[2026-06-24] update | S1 tracker checklist refreshed: DSA goal marked done (10 done Wed, 8/10 on time, 4 days early); daily plan Mon–Wed boxes checked (all 10 problems pulled forward to Wed); added Redis + DDIA Ch6 to weekly goals. Pending today: start SD URL Shortener / Redis intro.

[2026-06-28] update | Redis: reformatted System Design/Databases/Redis.md (frontmatter, headings, RESP summary table, fixed typos, added Interview Trade-Offs + persistence RDB/AOF). Created build-your-own-redis project dir (README staged roadmap + prereqs, python/server.py working Stage 1-4 selectors scaffold [py_compile OK], go/main.go reference stub, .gitignore). Language pick: Python (selectors = Redis event-loop model; Go goroutine-per-conn is opposite, off-plan). Indexed both pages (SD 23→25, page_count 130→132).

[2026-06-28] update | build-your-own-redis: removed code stubs (python/server.py, go/main.go, .gitignore) per request — user writes code. README converted to pure topic checklist (prereqs + 6 staged learning goals, concepts only, no code). Index summary updated.

[2026-07-02] update | Rate Limiter: formatted 8-page handwritten notes into System Design/Problem Designs/Rate Limiter.md (rewrote thin March skeleton). Added 5 algorithms (token/leaking/fixed-window/sliding-log/sliding-counter) w/ pros-cons + comparison table, 3 Mermaid diagrams (token bucket, leaking bucket, HLD), Redis INCR/EXPIRE, Lyft/Envoy rule format, rate-limit headers, distributed race condition (Lua/sorted-set) + sync (centralized store vs sticky session), perf optimization, hard vs soft, L7 vs L3 + OSI. Checklist: Week 06 S2 SD problem Rate Limiter marked done; Job Switch Hub SD table Rate Limiter ✅; URL Shortener noted deferred/pending. Page already indexed (no count change).

[2026-07-02] scorecard+retro | W05/S1 CLOSED (late, Thu): 8/10 on time (fail 621 formula, 355 merge), both hards solo, PASS STRONG. Retro written to Week 05 tracker. Coach Hub: W05 summary added, pointer → W06/S2, past-week table row added. Job Switch Hub: W5 DSA ✅ checked.

[2026-07-02] update | SD re-sequenced (user call): W06/S2 current-week SD = Design Key-Value Store + Consistent Hashing (added to Week 06 goals + SD tracker); W07/S3 SD = URL Shortener + Unique ID Generator (distributed) — swapped out Twitter Feed in Week 07. Job Switch Hub SD table re-sequenced to sprint weeks. Rate Limiter ✅ done.

[2026-07-04] update | Consistent Hashing: replaced thin note with full 12-section interview version (replication skip-walk, complexity/TreeMap upgrade, CoW concurrency, hot keys, CDN, sticky sessions, data migration cache-vs-persistent). Marked status: completed. Added focus markers (⭐ must-know / 🎯 focus / 💡 good-to-know) per section + focus-summary table. Index one-liner updated.

[2026-07-05] update | SD Theory Curriculum: reworked Progress from section-level to topic-level tracker. Marked confirmed-done topics ✅ w/ note pointers — Consistent Hashing ([[Consistent Hashing]]) + API Rate Limiter ([[Rate Limiter]]), plus existing Introduction + Databases. Flagged 📝 note-exists (review-pending, NOT done) for Caching, Message Queues/Kafka, URL Shorteners to avoid over-claiming. Rest ⬜ pending. Added status legend + Section Notes entries for Consistent Hashing + Rate Limiter.

[2026-07-05] create | Interview KB: Go Concurrency (Senior) — 18-entry deep-dive (GMP scheduler, chan internals, select, context, sync pkg, atomics, worker pool, fan-in/out, errgroup, race detector, memory model, leaks, 1.22 loop vars) + quick-recall table + 5 drills. Interview KB index: page 1 marked done.

[2026-07-05] ingest | Interview KB: Go Idioms & Interfaces page (Senior)

[2026-07-05] ingest | Interview KB: HLD Classic Set page (Staff)

[2026-07-05] ingest | Interview KB: HLD Scale Set page (Staff)

[2026-07-05] ingest | Interview KB: LLD OOD Set page (Staff)

[2026-07-05] ingest | Interview KB: Design a URL Shortener page (Staff, deep rewrite replacing crammed HLD Classic Set)

[2026-07-05] ingest | Interview KB: Rate Limiter page (Staff, written directly Sonnet, folded in existing Problem Designs note)

[2026-07-05] lint | Interview KB: archived HLD Classic Set / HLD Scale Set / LLD OOD Set (crammed multi-problem pages, superseded by per-problem deep pages) to _archive/, reordered index.md progress tracker to per-problem structure

[2026-07-06] ingest | Interview KB: SD Fundamentals wave — 9 small per-topic pages (Core Vocabulary & Quality Pillars, CAP Theorem & Consistency Models, Latency vs Throughput, Load Balancing, Proxies, Database Replication & Sharding, Single Point of Failure, Service Discovery & Health Checks, SD Interview Framework & Pitfalls), folded in from System Design Basics.md, Distributed Systems Concepts.md, Senior SD Mindset.md. Superseded an initial single-mega-page draft per user direction (topic-wise small pages, not one deep page).

[2026-07-06] ingest | Interview KB: 4 more SD Fundamentals pages (SQL vs NoSQL, CDN, Communication Protocols, Blob & Object Storage) — building blocks reused across 3+ upcoming HLD problems, added to Concepts/ per established small-page pattern. Python code examples included where concrete (long-polling loop, WebSocket handler, boto3 pre-signed URL).

[2026-07-06] ingest | Interview KB: pulled 5 more concept pages into KB for standalone completeness (Caching & Redis, Consistent Hashing, Message Queues & Kafka, API Design Principles, API Gateway). First 4 adapted from existing deep vault notes (System Design/Concepts/) with cross-links redirected to stay inside the KB. API Gateway was a stub in the original vault note (placeholder "add as you study") — written fresh at full depth (routing/auth/BFF pattern/gateway-vs-mesh distinction/Python routing skeleton). Fundamentals wave now 18 pages total, all inside Interview KB/System Design/Concepts/.

[2026-07-06] ingest | Interview KB: LLD wave started, reprioritized ahead of remaining HLD problems per user direction. Added 3 LLD foundation pages (LLD Interview Approach, SOLID Principles, OOD Design Patterns) folded from existing System Design/Low Level Design.md vault note (SOLID + Strategy/Observer/Factory/Decorator/Singleton patterns), all Python, standalone in KB. Wrote first LLD problem page: Parking Lot (Interview KB/System Design/LLD/Parking Lot.md) — class diagram, Strategy pattern for pricing, concurrency (lock granularity trade-off), edge cases, extensibility (EV charging, reserved spots), folding in existing Python code from the vault note and deepening it.

[2026-07-06] ingest | Interview KB: LLD Elevator System page (Interview KB/System Design/LLD/Elevator System.md) — SCAN/LOOK algorithm, State + Strategy patterns, elevator selection strategies table (nearest-idle/least-busy/zone-based/destination-dispatch), concurrency notes, edge cases, drills.

[2026-07-06] ingest | Interview KB: LLD LRU Cache page (Interview KB/System Design/LLD/LRU Cache.md) — OrderedDict fast-path + from-scratch doubly-linked-list + hash map implementation, sentinel-node technique, thread safety (lock then sharding follow-up), LFU variant, drills.

[2026-07-06] ingest | Interview KB: LLD Rate Limiter (OOD) page (Interview KB/System Design/LLD/Rate Limiter (OOD).md) — class-design/token-bucket version distinct from HLD Rate Limiter, Strategy pattern (TokenBucket/SlidingWindowLog), time.monotonic() rationale, per-client lock granularity, memory-bound extensibility, drills contrasting with distributed HLD version.

[2026-07-06] ingest | Interview KB: LLD Splitwise page (Interview KB/System Design/LLD/Splitwise.md) — Strategy pattern for split types (equal/exact/percent), greedy debt-simplification algorithm (max-debtor/max-creditor pairing) with mermaid flowchart, floating-point balance edge case, extensibility (new split types, group scoping, partial settlement), drills.

[2026-07-06] ingest | Interview KB: LLD Board Game page (Interview KB/System Design/LLD/Board Game.md) — naive O(N) win-check then O(1) running-sum optimization (the actual interview-scoring generalization), K-in-a-row extension for Gomoku-style boards, edge cases, extensibility (3+ players, Connect Four gravity), drills.

[2026-07-06] ingest | Interview KB: LLD Vending Machine page (Interview KB/System Design/LLD/Vending Machine.md) — canonical State pattern problem, state diagram + class diagram, full implementation vs rejected boolean-flags alternative shown side by side, concurrency (per-machine lock, stock check-then-act race), edge cases (exact change, mid-transaction cancel), extensibility (maintenance state, card payment, multi-item basket), drills.

[2026-07-06] ingest | Interview KB: LLD Notification System (OOD) page (Interview KB/System Design/LLD/Notification System (OOD).md) — Factory + Observer + implicit Strategy patterns, per-channel failure isolation via try/except, concurrency (lazy channel-dict race, async observer foreshadowing HLD version), extensibility (new channel registration, retry/rate-limit decorators), drills bridging to the HLD/Kafka version. LLD wave now complete: 8/8 problems + 3 foundation pages, all Python, all standalone in KB.

[2026-07-06] ingest | Interview KB: HLD News Feed/Twitter Feed page (Interview KB/System Design/HLD/News Feed - Twitter Feed.md) — hybrid fan-out (push for regular users, pull for celebrities) with Python merge logic, Snowflake time-sortable ID generator, capacity math (5.8K writes/sec, 325K reads/sec), Cassandra partition/cluster key reasoning, like-count batching, Staff lens on celebrity-case failure mode, 5 drills. Resumed HLD wave after LLD wave completed.

[2026-07-06] ingest | Interview KB: HLD WhatsApp Chat page (Interview KB/System Design/HLD/WhatsApp Chat.md) — WebSocket cross-server routing via Redis pub/sub registry with Python code, TTL-based crash self-healing, idempotent message handling (client_msg_id), Snowflake ordering + HLC clock-skew mitigation, presence heartbeat logic, group messaging fan-out size tiers, media pre-signed-upload flow, 5 drills.

[2026-07-06] ingest | Interview KB: HLD Notification System page (Interview KB/System Design/HLD/Notification System.md) — original vault note was a stub (placeholder "add as you study"), written fresh at full depth: priority-tier queue isolation (the Staff-lens failure mode), Python idempotency/dedup logic, exponential backoff + DLQ, rate limiting/digesting for frequency capping, quiet hours logic, cross-linked to LLD Notification System (OOD) companion page. 3/14 remaining HLD problems done this session (News Feed, WhatsApp, Notification System).

[2026-07-06] ingest | Interview KB: HLD Typeahead/Autocomplete page (Interview KB/System Design/HLD/Typeahead - Autocomplete.md) — no existing vault note, written from scratch. Trie with cached top-K per node (Python implementation, the core interview-scoring insight), batch trie-rebuild vs real-time contention trade-off, prefix-range sharding, personalization merge logic, 5 drills. 4/14 remaining HLD problems done this session.

[2026-07-06] ingest | Interview KB: HLD Web Crawler page (Interview KB/System Design/HLD/Web Crawler.md) — no existing vault note, written from scratch. Two-tier URL frontier (front/back queues, Mercator-style), Python Bloom filter implementation with false-positive/negative asymmetry explained, per-host politeness as structural queue design, SimHash near-dup mention, crawler-trap mitigation, 5 drills. 5/14 remaining HLD problems done this session.

[2026-07-06] ingest | Interview KB: HLD Distributed Key-Value Store page (Interview KB/System Design/HLD/Distributed Key-Value Store.md) — original vault note was empty, written from scratch (Dynamo-style). N/W/R quorum model with Python satisfaction check, vector clock conflict resolution with Python comparison logic (before/after/concurrent), hinted handoff + read repair + anti-entropy (cross-linked to existing Consistent Hashing Merkle-tree section rather than duplicating), gossip-based membership, 5 drills. 6/14 remaining HLD problems done this session.

[2026-07-06] ingest | Interview KB: HLD Payment System page (Interview KB/System Design/HLD/Payment System.md) — no existing vault note, written from scratch. Python idempotency-key reservation logic (atomic reserve-before-charge), append-only double-entry ledger vs mutable balance column, materialized balance view (CQRS), CP-not-AP framing as the canonical exception to eventual consistency, Saga pattern mention for cross-service atomicity, 5 drills. 7/14 remaining HLD problems done this session — halfway through the remaining HLD queue.

[2026-07-06] ingest | Interview KB: HLD Uber/Ride-Sharing page (Interview KB/System Design/HLD/Uber - Ride Sharing.md) — full geospatial deep dive: Python Geohash encoder, H3 hexagonal grid rationale (real Uber production choice), Redis GEOADD/GEOSEARCH, atomic driver-reservation race (same pattern as Parking Lot), surge pricing reusing matching's spatial index, 5 drills. Preserved personal war-story link to HCLTech geospatial tracking engine (250+ concurrent drivers, WebSocket, 4-5s updates) — cross-referenced for the STAR script. 8/14 remaining HLD problems done this session.

[2026-07-06] ingest | Interview KB: HLD YouTube/Video Streaming page (Interview KB/System Design/HLD/YouTube - Video Streaming.md) — DAG-based parallel transcoding pipeline, Python ABR player (sliding-window bandwidth estimation + safety margin bitrate selection), CDN caching key design (video_id+quality+chunk), view-count batching parallel to News Feed's like counts, live-vs-VOD distinction, 5 drills. 9/14 remaining HLD problems done this session.

[2026-07-06] ingest | Interview KB: HLD Dropbox/File Storage & Sync page (Interview KB/System Design/HLD/Dropbox - File Storage & Sync.md) — no existing vault note, written from scratch. Python content-addressed chunking (hash-based dedup + resumable upload), fixed-size vs content-defined chunking trade-off, versioning-via-shared-chunks metadata model with ref_count, conflicted-copy resolution strategy (Dropbox's real approach), 5 drills. 10/14 remaining HLD problems done this session.

[2026-07-06] ingest | Interview KB: HLD Ad-Click Aggregation page (Interview KB/System Design/HLD/Ad-Click Aggregation.md) — no existing vault note, written from scratch. Python windowed aggregator with event-time vs processing-time bucketing (the core interview-scoring insight), tunable grace period as explicit accuracy/latency trade-off, tumbling/sliding/session window taxonomy, dedup + fraud heuristics layered separately, exactly-once-via-idempotent-writes, raw-event-retention-enables-backfill, 5 drills. 11/14 remaining HLD problems done this session.

[2026-07-06] ingest | Interview KB: HLD Distributed Job Scheduler page (Interview KB/System Design/HLD/Distributed Job Scheduler.md) — no existing vault note, written from scratch. Python leased atomic-claim logic (same check-then-act race pattern as Parking Lot), exponential backoff + DLQ, idempotency-is-the-job's-responsibility framing, long-running-job lease-heartbeat fix, thundering-herd jitter, 5 drills. 12/14 remaining HLD problems done this session.

[2026-07-06] ingest | Interview KB: HLD Google Docs page (Interview KB/System Design/HLD/Google Docs.md) — folded in the best-quality existing vault note found this session (already covered OT vs CRDT, architecture, cursor presence, persistence, scaling well), extended with Python OT transform function and CRDT RGA structure with deterministic tie-breaking, 5 drills. 13/14 remaining HLD problems done this session — one left (Distributed Message Queue).

[2026-07-06] ingest | Interview KB: HLD Distributed Message Queue page (Interview KB/System Design/HLD/Distributed Message Queue.md) — the Kafka-internals problem (not "use a queue" but "build one"), Python append-only log + key-based partition hashing + consumer-offset tracking with idempotent-commit-after-processing, acks=0/1/all durability trade-off, ISR leader election, log compaction, 5 drills. HLD wave now COMPLETE: 16/16 problems (URL Shortener + Rate Limiter done earlier, 14 more this session: News Feed, WhatsApp, Notification System, Typeahead, Web Crawler, Distributed KV Store, Payment System, Uber, YouTube, Dropbox, Ad-Click Aggregation, Job Scheduler, Google Docs, Distributed Message Queue). Interview KB System Design section (Concepts + HLD + LLD) fully standalone and complete.

[2026-07-06] lint | Interview KB: removed _archive/ (HLD Classic Set, HLD Scale Set, LLD OOD Set) — confirmed all 24 problems these covered now have individual per-problem replacement pages (16 HLD + 8 LLD, all complete). Nothing unique lost; recoverable via git history if ever needed.

[2026-07-06] ingest | Interview KB: Python Core wave started (Wave 3). Two pages: Identity, Mutability & Memory (is vs ==, mutable default arg gotcha, shallow/deep copy, [[]]*n reference trap, refcounting+GC) and Python Scoping & Closures (LEGB, closures, late-binding loop gotcha with fix). Folded in from existing Python Interview Questions.md / Python Programming.md, deepened with why-explanations and drills.

[2026-07-06] ingest | Interview KB: Python Decorators page (Interview KB/Backend/Python/Python Decorators.md) — 3-level nested pattern (factory/decorator/wrapper), functools.wraps rationale tied to real Flask route-collision failure mode, dependency-injection-via-decorator-arguments example, folded in from existing Decorators.md vault note plus Python Interview Questions.md.

[2026-07-06] ingest | Interview KB: Python Generators & Iterators page — lazy evaluation vs list materialization, yield/StopIteration mechanics, generator expressions, class-based vs @contextmanager-generator context managers with the shared pause/resume mechanism explained, 3 drills.

[2026-07-06] ingest | Interview KB: Python OOP & Data Model page — four pillars precisely distinguished, name-mangling honesty on encapsulation, MRO/super() under multiple inheritance, duck typing, dunder methods table + str vs repr, @property/@classmethod/@staticmethod, @dataclass mutable-default fix tied back to Identity/Mutability page, 3 drills. Folded in from Object Oriented Programming.md + Python Interview Questions.md.

[2026-07-06] ingest | Interview KB: Python Concurrency page — GIL decision framework (threading/multiprocessing/asyncio table with "why" per tool), run_in_executor thread-pool-vs-process-pool distinction, cross-linked to deep Asyncio mechanics page rather than duplicating, 3 drills.

[2026-07-06] ingest | Interview KB: Python Gotchas Cheat Sheet page — fast-recall table cross-linking to full explanations on their dedicated pages, time-complexity table, DSA built-ins reference. Python Core wave complete: 7/7 pages (Identity/Mutability/Memory, Scoping & Closures, Decorators, Generators & Iterators, OOP & Data Model, Concurrency, Gotchas Cheat Sheet).

[2026-07-06] ingest | Interview KB: Backend Libraries wave complete (Asyncio, Pydantic, FastAPI, Python Logging) — all folded in from already-excellent production-quality vault notes, adapted frontmatter + standalone cross-links, added interview-drills sections to match KB convention. Notable ties: FastAPI's response_model framed as an output-shape security boundary, BackgroundTasks contrasted with real durable queues, lazy %s-logging vs eager f-string evaluation. Wave 3 (Python) now COMPLETE: 11/11 pages (7 core language + 4 libraries).

[2026-07-06] ingest | Interview KB: Wave 4 (Databases) started. First 7 SQL pages folded in from already-excellent vault notes (SQL/ folder, 16 files) with light adaptation (frontmatter, standalone cross-links, drills added): SQL Basics & Query Fundamentals, SQL JOINs, SQL Aggregates & GROUP BY, SQL Window Functions, SQL Subqueries & CTEs, SQL Indexes & Query Performance, SQL Transactions & ACID. 8 more SQL pages planned (String/Date, UNION, Pivoting, Constraints/DDL, Views/Procs/Triggers, Security, Classic Interview Problems, Style/Best Practices) then Postgres Internals, Redis (cross-ref existing deep Caching & Redis KB page), MongoDB.

[2026-07-06] ingest | Interview KB: SQL wave complete — 15/15 pages, all in Interview KB/Backend/SQL/. Remaining 8 folded from vault (String & Date Functions, UNION & Set Operations, Pivoting, Constraints & Schema DDL [combined from Constraints.md + DDL DML DCL TCL.md], Views/Procedures/Triggers, Security & Injection Defense, Classic Interview Problems [from SQL for Interviews.md's problem set + Instacart case study], Style & Best Practices [from Clean SQL.md]). Next: Postgres Internals, Redis (cross-ref existing deep Caching & Redis KB page rather than duplicate), MongoDB.

[2026-07-06] ingest | Interview KB: Wave 4 (Databases) COMPLETE. Postgres Internals page (Interview KB/Backend/SQL/Postgres Internals.md) — xmin/xmax tuple versioning mechanism, VACUUM/bloat, TOAST, WAL + streaming/logical replication, PgBouncer pooling modes, avoids duplicating SQL Transactions & ACID / SQL Indexes pages. MongoDB page (Interview KB/Backend/SQL/MongoDB.md) — written from scratch: embed-vs-reference modeling decision, sharding hot-shard risk tied to Database Replication & Sharding page's general pattern, aggregation pipeline, multi-doc transactions framed as a modeling smell, honest MongoDB-vs-Postgres+JSONB framing. Redis: no separate page written — existing Caching & Redis Concepts page (Wave 0) already covers Redis in full depth; index cross-references it instead of duplicating. Wave 4 total: 15 SQL pages + Postgres Internals + MongoDB = 17 new pages.

[2026-07-06] ingest | Interview KB: Wave 5 (Frontend, final wave) started. 5 JS pages folded from excellent vault notes (JavaScript.md, JavaScript Interview Questions.md, JavaScript Function Code Examples for Interviews.md, JavaScript DOM Notes.md): JS Core Language & Data Types, JS Closures & Async, JS Prototype & OOP, JS Machine Coding Patterns (debounce/throttle/curry/promise-polyfills/array-polyfills/deep-clone with cycle handling), JS DOM & Events. All in Interview KB/Frontend/JS/. Next: React wave (4 pages) using React.md + Frontend System Design.md's real machine-coding problems (Typeahead/Infinite Scroll/WebSocket Dashboard/Forms), then HTML & CSS.

[2026-07-06] ingest | Interview KB: React wave complete (4 pages) — React Hooks Deep-Dive, React Patterns & State Management, React Performance & Reconciliation, React Machine-Coding (Frontend System Design). Last page folds in Frontend System Design.md's 4 real machine-coding problems (Typeahead w/ AbortController race-condition fix, Infinite Scroll w/ cursor pagination, WebSocket Dashboard w/ shared connection + backoff, Complex Form), cross-linked to Typeahead/News Feed/WhatsApp HLD pages already in KB. One page left in entire KB build: HTML & CSS Fundamentals.

[2026-07-06] ingest | Interview KB: HTML & CSS Fundamentals page (Interview KB/Frontend/HTML & CSS Fundamentals.md) — no existing deep vault note (only a question list), written from scratch: semantic HTML/a11y rationale, script loading strategies, box model, positioning gotcha (absolute-positioning-context fix), Flexbox vs Grid decision rule, specificity tuple, z-index/stacking-context gotcha, CSS custom properties vs Sass variables for runtime theming, 3 drills.

[2026-07-06] lint | ENTIRE INTERVIEW KB BUILD COMPLETE. Final tally: Go (2) + SD Concepts (18) + SD HLD problems (16) + SD LLD (8 problems + 3 foundations) + Python (7 core + 4 libraries) + Databases (15 SQL + Postgres Internals + MongoDB, Redis cross-ref) + Frontend (5 JS + 4 React + 1 HTML/CSS) = 85 pages total, all standalone within Interview KB/, all cross-linked, Python/Go code included wherever a topic reduces to concretely codeable logic. index.md progress tracker fully updated to reflect completion across all 5 waves.

[2026-07-06] lint | Interview KB: fixed 4 empty scaffolded folders. Moved Postgres Internals.md and MongoDB.md from Backend/SQL/ to their own Backend/Postgres/ and Backend/MongoDB/ folders (correct scaffold location); moved HTML & CSS Fundamentals.md from Frontend/ root into Frontend/HTML-CSS/. Added a README pointer in Backend/Redis/ explaining Redis content intentionally lives in System Design/Concepts/Caching & Redis.md rather than duplicating it there. Zero empty folders remain in Interview KB/.

[2026-07-06] lint | Interview KB: fixed the Go wave (written earliest in the session via Fable, before later conventions locked in). Go Concurrency.md had 7 broken wikilinks to [[Backend Libraries]] and [[SD Concept Catalog]] — both pages never existed under those names (superseded early on: Backend Libraries became 4 separate pages Asyncio/Pydantic/FastAPI/Python Logging, SD Concept Catalog's content was absorbed into the 18 individual Concepts pages). Each fixed with a contextually correct target: back-pressure entry -> [[Message Queues & Kafka]]; context/DI entry -> [[SOLID Principles]]; worker-pool-sizing entry -> [[Python Concurrency]] (cross-language parallel); remaining entries (goroutine leaks/pprof/goleak, errgroup, scheduler) had the broken link dropped since no KB page covers those Go-specific tools. Also added missing type:/difficulty: frontmatter fields to both Go pages to match the convention established later in the session. Verified zero remaining broken [[Backend Libraries]]/[[SD Concept Catalog]] references anywhere in Interview KB/.

[2026-07-06] lint | Interview KB: reformatted the Go wave to match the KB's small-per-topic-page convention used everywhere else. The original 2 mega-pages (Go Concurrency 560 lines/18 entries, Go Idioms & Interfaces 607 lines/18 entries — written first in the session via Fable, before this convention was established) were split into 14 small pages: Go Goroutines & Scheduling, Go Channels, Go Context & Cancellation, Go Sync Primitives, Go Concurrency Patterns, Go Race Conditions & Memory Model, Go Goroutine Leaks, Go Interface Internals, Go Type Assertions & Generics, Go Embedding & Dependency Injection, Go Error Handling, Go Data Structure Internals, Go Receivers & Control Flow, Go Zero Values Enums & Options Pattern. Content preserved near-verbatim (already excellent, senior-level depth) — reformatted headers, redistributed shared quick-recall/drill sections to their relevant new page, fixed internal "#N" cross-references to point to the correct sibling page or drop where no longer applicable. Old mega-pages deleted (not archived — fully superseded, zero content lost). KB total now 97 pages (up from 85; net +12 after removing the 2 originals and adding 14).

[2026-07-06] lint | Interview KB: rebuilt index.md from scratch. Removed all progress-tracker noise (wave/checkpoint commentary, "done" notes, format-note paragraphs, stale duplicate Study Order text) accumulated during the build. Replaced with a clean navigational index matching vault convention (one-line summary per page), reordered into user's stated priority: SD Concepts (18) -> SD HLD (16) -> SD LLD (11: 3 foundations + 8 problems) -> Python (11) -> Databases (17) -> Go (14) -> Frontend (10) = 97 pages verified against actual files on disk. Every link confirmed to resolve to a real file before writing.

[2026-07-06] lint | Interview KB: index.md was found deleted from disk (git showed it staged as deleted, not by my prior edits) — recreated with the same clean rebuilt content (97 pages, ordered SD Concepts -> HLD -> LLD -> Python -> Databases -> Go -> Frontend, verified links, one-line summaries).

[2026-07-06] ingest | Interview KB: added Merkle Trees page (System Design/Concepts/Merkle Trees.md) — was previously only a dangling cross-reference between Consistent Hashing and Distributed Key-Value Store (each pointed at the other for "the full mechanism," but neither actually explained it). New page covers: hash-tree structure, Python build + divergence-detection code, why root-hash comparison works, O(log n) pruning logic, real usage (Cassandra nodetool repair, Git, Bitcoin), 3 drills. Fixed both existing pages' Merkle mentions to point at this new page instead of each other. Added to index.md SD Concepts section (19 pages now); KB total 98 pages.

[2026-07-06] ingest | Interview KB: added §9 Read Path & Write Path to Distributed Key-Value Store.md — step-by-step PUT/GET walkthroughs with Python code, tying together consistent hashing (§5), quorum (§6), vector clocks (§7), and hinted handoff/read repair (§10) into one concrete sequence instead of leaving them as separate concepts. Renumbered subsequent sections (Handling Failures ->10, Deep Dives ->11, Drills ->12), fixed internal §-references, added a 6th drill question about the write-returns-after-W-not-N timing.

[2026-07-06] ingest | Interview KB: added §10 Per-Node Storage Engine to Distributed Key-Value Store.md, citing the real Cassandra/Dynamo LSM-tree architecture (commit log -> memtable -> SSTable flush for writes; memtable check -> Bloom filter -> indexed SSTable lookup for reads; compaction) per user-provided reference material. Cross-linked to Postgres Internals (WAL parallel) and Web Crawler (Bloom filter mechanism reused). Renumbered Handling Failures->11, Deep Dives->12, Drills->13, fixed all internal section references, added a 7th drill question. Page now has 13 sections total, going from coordinator-level fan-out (§9) down to what a single node actually does with the write/read (§10) — the full depth stack an interviewer would push toward.

[2026-07-06] ingest | Interview KB: added Design a Unique ID Generator.md (System Design/HLD/) — user found the gap ("where is [this problem]") after Snowflake IDs had only ever been covered inline within URL Shortener/News Feed/WhatsApp Chat, never as its own standalone canonical page despite being one of the most commonly asked standalone HLD problems. New page covers all 4 standard approaches (multi-master auto-increment, UUID, ticket server, Snowflake), full Python Snowflake implementation with bit-budget reasoning, clock-skew-backward edge case handling, comparison table, 5 drills. Cross-linked from and to the 3 existing pages that touch Snowflake, added to index.md as 17th HLD page. KB total now 99 pages.

[2026-07-07] update | Distributed Key-Value Store.md: expanded gossip failure-detection mechanism in §8 beyond "propagation" — added concrete SWIM-style protocol (direct probe -> indirect probe via k random peers -> suspicion state with refutation window -> gossip-piggybacked dissemination), plus why this beats naive central heartbeats (O(1) msgs/node/round, O(log N) detection time). Note: [[Service Discovery & Health Checks]] cross-link target does not exist yet as a page — remains a dangling reference.

[2026-07-12] update | Distributed Key-Value Store.md: added heartbeat-counter gossip as the baseline failure-detection model in §8 (membership list of (id, counter, local-recv-time); increment-own-counter, gossip-full-list-to-random-peers, merge-by-max-counter, offline-on-stale-counter) with Python merge/detect snippets. Explained the local-receive-time vs remote-clock subtlety (no synced clocks needed, sidesteps §7 LWW skew) and why counters alone give unwinnable false-positive-vs-detection-speed tension — motivating SWIM. Added production usage table: Cassandra/ScyllaDB Phi Accrual, original Dynamo heartbeat gossip, Consul/Nomad/Serf SWIM (memberlist), Redis Cluster PFAIL->FAIL quorum. User-requested after comparing their heartbeat-counter description against the file (which jumped straight to SWIM, skipping the baseline).

[2026-07-16] ingest | Back-of-Envelope Estimation.md: new SD Concepts page (est. numbers to memorize — powers of 2, time-to-seconds, Jeff Dean latency ladder; the 5-quantity estimation recipe DAU→QPS→ratio→storage→bandwidth→cache; worked Twitter-scale feed example deriving shard/cache/fan-out decisions from the numbers; single-node sanity-check ceilings; 2 interview drills). Cross-linked to SD Framework, Core Vocabulary, Latency vs Throughput, Caching & Redis, Replication & Sharding, Twitter Feed. index.md: Concepts 19→20, total 99→100.