Skip to content
DSA Grind
All 26 sections

Step-by-Step Study Plan — 20 Weeks to MAANG-Ready

NoteUpdated
On this page

Tailored for: Sanjay Naik — 3.5 yrs Java backend, microservices + GenAI healthcare exp Goal: Senior SWE / SDE-II / L4 / E4 offer Start: 2026-05-20 → Target onsites: October–November 2026

Phase 0 — Pre-flight (Week 0, 3 days)

  • Set up LeetCode Premium (or use free + NeetCode + LCUS company tags spreadsheets)
  • Buy Hello Interview / Educative System Design subscription (1 month) — defer to M3
  • Set up an Anki deck (or use the pattern-guide markdown in ../ as your deck) for spaced review
  • Block 3 hrs/weekday + 6 hrs each Sat/Sun on calendar — non-negotiable
  • Create a tracker sheet: problem #, pattern, time taken, mistakes, review dates (D+1, D+3, D+7, D+21)
  • Take a baseline: solve 5 random LC mediums in 25 min each. Record success rate.

Phase 1 — DSA Foundations (Weeks 1–4)

You already have the parent DSA Study/ folder (../) with 20 pattern guides. Don’t re-write notes — solve.

Week Patterns Problems target Companion reading
1 Two Pointers, Sliding Window 16 (8+8) ../01-Two-Pointers/, 02-Sliding-Window/
2 Hashing / Frequency Maps, Fast & Slow Pointers, Merge Intervals 18 03-, 04-, 05- folders
3 Cyclic Sort, Linked List Reversal, BFS 18 06-, 07-, 08-
4 DFS, Top-K, Two Heaps 18 09-, 10-, 13-

End of M1 checkpoint:

  • Blind 75: 100% solved
  • 70 unique LC problems solved
  • Can recognize the pattern in < 2 minutes for 80% of new problems
  • 1 mock interview booked (Pramp) — record it, watch yourself

Phase 2 — Advanced DSA + LLD warm-up (Weeks 5–8)

Week Topics LC count LLD project
5 Backtracking / Subsets, Binary Search (modified) 16 Parking Lot (Java)
6 Dynamic Programming I (1D, knapsack) 12 Rate Limiter (you’ve built one — productize as LLD spec)
7 Dynamic Programming II (2D, LIS, edit distance) 12 LRU Cache + Splitwise
8 Graph: Union-Find, Topological Sort, Dijkstra 12 Logging Framework + Vending Machine

End of M2 checkpoint:

  • NeetCode 150: 100% solved
  • 130 unique LC problems
  • 5 LLD designs documented in 03-LLD/
  • Mock #2: medium LC in 25 min consistently

Why LLD early: it pays double — same pattern shows up in HLD database-design portion + you can answer “design X” questions concretely.

Phase 3 — System Design HLD (Weeks 9–14)

This is where you win vs other 3.5-yr candidates. Your Jio work (CDC, Kafka, LLM rate limiter, EHR pipelines) maps directly to L4 HLD bar.

Week 9: HLD fundamentals refresh

  • CAP, PACELC, consistent hashing, quorum, leader election
  • Load balancing (L4 vs L7), CDN, caching (cache-aside, write-through, write-back)
  • DB sharding strategies, replication (sync vs async), read replicas
  • Message queues — Kafka deep dive (you already use it — re-frame in interview terms)
  • Write notes in 02-System-Design/fundamentals.md

Week 10: Designs 1–3 (read pattern, then practice on whiteboard in 45 min)

  • Design URL shortener (TinyURL)
  • Design Twitter timeline
  • Design WhatsApp / Messenger

Week 11: Designs 4–6

  • Design Uber / Lyft (geo-hashing, S2)
  • Design YouTube / Netflix (video pipeline) — frame as your transcoding project
  • Design Instagram feed (push / pull / hybrid)

Week 12: Designs 7–9

  • Design Google Drive / Dropbox
  • Design distributed cache (Redis cluster)
  • Design rate limiter — your AI rate-limiter story goes here

Week 13: 2026 GenAI-infra designs (~50% of SD loops now)

  • Design LLM API serving (variable-latency, batching, KV cache, GPU pooling)
  • Design RAG system (vector DB, ingestion, retrieval, hybrid search)
  • Design “Design Cursor” / “Design Claude Code-style agent backend”
  • Frame your Clinical AI Summarization Engine as a real example

Week 14: Healthcare-domain SD (for healthtech rounds — Verily, Oscar, Cigna, Aetna)

  • Design HIPAA-compliant EHR sync (you’ve built this — write it up)
  • Design FHIR resource server with eventual consistency
  • Design real-time medical-alerts pipeline (Debezium → Kafka → ML scoring)

End of M3 checkpoint:

  • 13 HLD writeups committed in 02-System-Design/
  • 1 mock SD interview (Exponent or interviewing.io)
  • Start applying to companies — referrals first, then direct
  • Resume rewritten per RESUME-VALIDATION.md

Phase 4 — Behavioral + Polish (Weeks 15–18)

Week 15: Write 12 STAR stories

Map each to ≥ 2 Amazon LPs. Templates and pre-filled drafts in 05-Behavioral/.

Recommended story portfolio (you have raw material for all of these):

  1. Customer Obsession — AI Rate Limiter saved 40% LLM cost ⟶ customer-facing latency improvement
  2. Ownership — End-to-end ePRO microservice across 5 services
  3. Invent and Simplify — AI-Driven Development Framework (60% dev time cut)
  4. Are Right, A Lot — Hibernate optimization 2.5s → 500ms (data-driven decision)
  5. Learn and Be Curious — Built Kong + Konga POC from scratch
  6. Hire and Develop — KT sessions, 40% sprint velocity gain
  7. Insist on the Highest Standards — 200+ SonarQube vulnerabilities mitigated
  8. Think Big — Clinical AI Summarization Engine across longitudinal EHR
  9. Bias for Action — CDC microservice with sub-3s sync, shipped in N weeks
  10. Frugality — Cost optimization via LLM rate limiter
  11. Earn Trust — Code reviews + technical governance role
  12. Dive Deep — Hibernate fetch-cycle elimination root-cause
  13. Have Backbone; Disagree and Commit — pick a real disagreement (architecture call)
  14. Deliver Results — 80+ partner hospitals on Medical Reporting Engine
  15. Strive to Be Earth’s Best Employer — mentorship / inclusion (lower stakes; skip if no story)
  16. Success and Scale Bring Broad Responsibility — HIPAA compliance & security work

Week 16: Rehearse + record (Loom yourself, listen back, tighten to ≤ 3 min each)

Week 17: 4 full-loop mocks (1 Meta-style, 1 Google-style, 1 Amazon-style, 1 healthtech)

Week 18: Iterate on weak signals from mocks; re-do 20% problems you got wrong

Phase 5 — Application & Onsites (Weeks 19–20+)

  • Recruiter reach-outs / referrals: start in Week 12 (Phase 3), not now
  • Apply by tier: Microsoft/Atlassian/Walmart-Labs/Adobe (warm-up) → Amazon/Meta/Google (target) → Apple/Netflix (stretch)
  • 1 problem/day to stay sharp; rotate companies’ tagged Qs every 3 days
  • After each onsite: write a post-mortem in 99-debrief/ (create when needed)

What you can SKIP (saves 60+ hours)

  • ❌ Hard DP problems (matrix chain multiplication, palindrome partitioning II) — < 5% of L4 loops
  • ❌ Segment trees, Fenwick trees — only Google occasionally, low ROI
  • ❌ Competitive-programming tricks (manacher, KMP) — almost never asked
  • ❌ Frontend / CSS — you’re backend; redirect to Spring/Kafka instead
  • ❌ Re-writing your DSA notes — they exist already; solve, don’t re-document

Risks (call out now, mitigate early)

Risk Mitigation
Burnout from 3 hrs/day for 5 months Saturday OFF every 4th week. Plan it.
Coding speed under pressure Force 25-min timer from week 1. Don’t pause it.
Behavioral “tell me about yourself” rambling Write & memorize ≤ 90-sec elevator pitch, week 15
Last-min anxiety on system design Do mock #3 by week 14 — not week 18
Healthcare-only stories sounding niche For Meta/Google, lead with technical complexity not domain

Tracking template (copy into a sheet)

Date | Problem | Pattern | Time | Pass? | Mistake | Next review

Open 01-DSA/top-questions.md next and start ticking.