Interview Notebook

Notes for Adarsh Vishwakarma. Java Live Code. Job 10454435 still has no public IE that names a live-round question. Unnamed stays unnamed. R1 (13 Aug 2026), R2 (18 Aug 2026), and R3 are done. OA is not a round. Use the Round 4 prep book for the leftover hour. The chapters below are companions (DSA, LLD, stack). Ignore Round 2 = 14 Aug in older research files.

ConvBI / Ylogx LP drill (expand/collapse): ConvBI-LP-Interview.html — nested follow-ups on resume claims, Ownership/Frugality/Dive Deep primaries, reliability 99.9%, Java retries labeled understand vs shipped. ConvBI only for now.

HTML is the book. Every later-loop answer needed for rehearsal is in the chapters below: full Java, spoken STAR, project mechanisms, production failure classes, technology trade-offs, and resume traps. The files under _answers/ are provenance and source dumps only; you do not need to open them to prepare. The existing study book, research bible, Answer-BIBLE, and question banks remain read-only companions.

Questions are labeled IE-asked Resume-derived Standard CS. A prompt from another candidate’s loop is not a Job 10454435 forecast.

How to use the hour

  1. Round 4 leftover hour → r4/index.html. Intro, stories, leftover prompts, projects, Java if they paste code. That book is candidate-facing (what they might ask / what you say), not an interviewer scoring sheet.
  2. If they start with “tell me about yourself” and you are already in the Round 4 book → r4 chapter 2. Companion intro: chapter 02.
  3. If they ask LP / behavioral / STAR → open Round 4 first: r4/09-lp-questions.html (unique LP question list; click a prompt to the answer), then the Q&A cards in r4/09-lp-answers-a.html and r4/09-lp-answers-b.html, r4/10-behavioral.html for non-LP behavioral, and r4/11-resume-mock.html if they work the PDF open. LP means STAR mapped to the 16 official Leadership Principles; behavioral is intro, why Amazon, live issues, and questions for them. Companions: chapter 21 for which question → which locked story → talking points, chapter 12 for the full spoken STAR, and Round 4 stories. Failure = Argus 73%. Deadline = no fake miss. Conflict = hybrid versus vector-only or RLS-in-database versus app-only. Open practice-not-yours only to recognize claims you must never speak as Adarsh. DO NOT SPEAK PRACTICE AS ADARSH.
  4. Fluency-as-DSA: if they paste a tree and the invite said Fluency, write the tree first: Distance K in chapter 05, as reported in LC 7623949. Intern LC 7643209 was an unnamed min-heap, not Distance K. The complete later-slot Java bank is in chapter 05; give the Fluency answer only if they still ask after the code.
  5. If they grind why-this-tech, live issues, Horizon, or the PDF itself, use the complete HTML chapters: technology why/why-not, Ylogx dev→QA→prod, ConvBI / Ylogx full codebase (modules, RLS, SQL RAG, deploy from the repo), Horizon, hackathons, and IEDC, and resume critique.
  6. If they stay on one intern or project → chapter 19 (timeline + Q&A) and Round 4 projects. Spoken STARs: Round 4 stories.
  7. If they point at a skills-line technology (Kubernetes, GraphQL, FAISS, Auth0, Java-as-intern, Redis-as-rate-limiter) — listed ≠ shipped — open chapter 20. Point to the evidenced hop. Do not fake EKS or cluster-owner theatre.
  8. If they want design → chapter 22 (end-to-end LLD + why this DS) and chapter 23 (HLD boxes; not UTA-default). Shorter locker/parking/rate-limiter rehearsal stays in chapter 06. Rate Limiter is one SDE I Second Technical (OA+3), not UTA two-DSA.
  9. If they paste DSA and the algorithm is what is unclear → DSA Notes booklet (two examples and two Java versions per algo, then a table into the named-IE cards). If you already know the algo and need the 30-second scan → chapter 25 then chapter 24. Full named-IE Java stays in chapter 05. Root Cheat-Sheet.md remains the older companion scan; do not treat it as a Job 10454435 list.
  10. If they open GitHub on InstaRecon: security-awareness demo, consent, no production attacks. Then Ylogx / IQVIA / Argus / StratifyLabs. Mismatches: chapter 17.
  11. If a new IE paste arrives → remap in chapter 18 before studying it as your third live. OA-as-R1, INTERN, US, FTC, SDE-2 HLD, YOE = not your format.

Chapters

  1. Lock and evidence — OA-as-R1 trap, what “their Round 3” means, research conflicts
  2. Intro and how to speak — 90s / 45s, knowing vs not-knowing
  3. What Round 3 can be — formats after two lives, 60-minute shapes, not a prediction
  4. Round 4, Bar Raiser, HM — what follows a third live
  5. DSA for later lives — Java, named IE plus pattern cards. Algorithm teaching (two examples, simple + interview Java) is the DSA Notes booklet
  6. OOD / LLD / DS-design — locker, parking, search, rate limiter labeled
  7. GenAI Fluency — when it is RAG talk, when it is Distance K
  8. Leadership Principles — R3/R4 probes, spread, spoken STAR
  9. Resume projects — IQVIA, Ylogx, Horizon, Argus, GiftedBooks, StratifyLabs
  10. Stack, Java, net, AWS, CNN — what is on the resume vs Standard CS
  11. What to prepare / what not to — highest-value list, no fake grind
  12. Genuine STAR — all 16 LPs, BR probes, why-chain, what not to say
  13. Practice STAR (not yours) — eight fake templates that get busted; do not speak as Adarsh
  14. Why this stack — Azure hybrid, LangGraph, RLS vs RBAC, Redis, YOLO, GStreamer, honest 10×
  15. Ylogx live path — blast radius, how RLS/RBAC, JOIN leak / wrong-tier cache / LLM on hot path, ECS/ALB
  16. Horizon, ERC, hackathons, IEDC — 17th/80+, stale occupancy, named prizes, no mentee count
  17. Resume-critique mock — imprecise %, skills vs bullets, GitHub mismatches, 20-second bullets
  18. New IEs remapped — their round vs your third live; not a 10454435 prediction
  19. Intern and project prep — timeline and Q&A for each intern and project; not a 10454435 prediction
  20. Resume stack defense — FACT vs listed-not-shipped vs MISSING; Tier 1/2/3 cards; Kubernetes honest (ECS shipped); not a 10454435 prediction
  21. LP question bank — Exponent / IGotAnOffer / LC maps onto locked resume stories + lean-in talking points; Class C vs IE-asked; not a 10454435 prediction
  22. LLD / DS-design answerbook — meeting rooms, backup/restore, ACL share, log parser, moving average, parking, vending, LMS, elevator, LRU, playlist mix, audio buffers, TUF bookstore/email/autocomplete/notification/LCA/Min Stack; DS tables + Java; not a 10454435 prediction
  23. HLD answerbook — URL shortener, rate limiter (count = 1), notifications, S3-like, ride/order, metrics/logs, Redis-like; not UTA-default; not a 10454435 prediction
  24. Named-question cheat sheet — live DSA families from the research bible + TUF + chapter 05: why this, brute, optimal, time and space; LP scan last in the same file; not a 10454435 prediction
  25. Clue → pattern cheat sheet — the contiguous / at-most-K / next-greater / top-K / maximize-minimum table with why the wording forces the pattern, why brute dies, Java DS, TC/SC; not a 10454435 prediction
  26. ConvBI / Ylogx full codebase — every Nest/FastAPI/React module, RLS and SQL-RAG algorithms, dashboards, Tally, ECS/GH Actions; leftover hour / possible R4; not a 10454435 prediction
  27. Antigravity Story Bank v2 — LP cards and project deep dives
  28. Bar Raiser resume metrics — metric proofs + technical flows (Horizon costmap, IQVIA RAG, GiftedBooks FAISS, LangSmith, StratifyLabs)
  29. Bar Raiser technicalities — every technical deep-dive prompt with mapped projects, real failure modes, trade-offs, and live-code-style Java fragments

Evidence and provenance files

The HTML chapters above are the complete rehearsal book. The markdown files are retained only so a claim can be traced back to its source: R3-R4-Question-Bible.md §18, the pasted-IE remap, the answer-source index, and the targeted miner output. They are not required reading and do not replace any spoken answer in HTML. The targeted miner was not a 500-host sweep; existing research and study-book files were not rewritten.

Resume metrics (do not invent extras)