Zerodha interviews reward ownership and working systems over pure LeetCode volume. The notes below synthesise public interview guides (SpaceComplexity, InterviewEra), community assignment repos, and medium-post style candidate reports into round tables and outcome notes - confirm every stage with your recruiter for your role.
Round-wise process at a glance
Round
Typical format
What clears you
Resume screening
Async review
Fintech/markets curiosity, side projects, readable Git history
OA / take-home
HackerRank 60–90 min and/or 90–120 min build
Correct, runnable solution + README
Technical × 1–2
45–60 min each
Defend design, trade-offs, production thinking
Culture / HR
30–45 min
Ownership, simplicity, alignment with bootstrapped model
Timelines of 2–6 weeks between stages are commonly reported - keep one owned project warm for deep follow-ups.
Candidate profile: Backend-leaning SWE applicant; composite of published Zerodha process guides.
Stage 1 - Take-home assignment (≈1 week window)
Build a small service from a product brief (reported themes: order handling, rate limiting, or data ingest + search UI)
Expected deliverables: working API or script, tests or smoke checks, README explaining setup and trade-offs
Git history and commit messages are part of the signal - not a single “final dump” commit
Tip: Document what you did not build and why; lean teams penalise mystery complexity
Stage 2 - Technical interview (45–60 min)
Walk through your assignment: schema, concurrency, error handling, idempotency
Follow-ups: “What breaks at 10k concurrent requests?” “How would you deploy and monitor this?”
Domain prompts: explain buy vs sell, limit vs market order at a high level; use Kite once before the call
Result: Advanced to culture round in accounts that cleared technical depth
Stage 3 - Culture fit (30–45 min)
Why Zerodha over a funded startup? Examples of end-to-end ownership
Comfort with small teams, direct feedback, and long-term product thinking
Takeaways
What decided it: A production-quality take-home (tests, README, honest scope cuts) defended on concurrency, idempotency, and deploy/monitor follow-ups, plus bootstrapped-culture ownership stories.
Common fail: Mystery complexity in the assignment; cannot explain buy/sell or limit vs market at a high level.
How to practise: Ship a small order/rate-limit service with Git history; rehearse “what breaks at 10k concurrent requests?”
Technical interview experience 2 - HackerRank OA + live machine coding
Technical interview experience 2
Candidate profile: Mid-level applicant pattern from SpaceComplexity-style write-ups.
Round 1 - Online assessment (60–90 min)
DSA and debugging on HackerRank or similar - arrays, strings, hash maps, basic graphs
Less emphasis on trick puzzles; clarity and edge cases matter
Round 2 - Machine coding (90–120 min, live or take-home)
Prompt example: implement a simple order book or rate limiter with clean interfaces
Discuss partial fills, cancellation, or token-bucket vs leaky-bucket verbally
Over-engineering (custom infra with no requirement) is a negative signal
Round 3 - System-leaning technical (45–60 min)
Order state machine: pending → open → partially filled → filled / cancelled / rejected
Margin impact at each transition; idempotency keys for retries
Stack awareness (Go for hot paths, Kafka/WebSockets conceptually) without trivia
Takeaways
What decided it: Clear OA edge cases, then a simple order book/rate limiter with partial fills discussed, plus drawing the pending→filled state machine before deep follow-ups.
Common fail: Over-engineering infra with no requirement; fuzzy order lifecycle under margin/idempotency probes.
How to practise: Draw the order state machine on paper; implement token-bucket verbally vs leaky-bucket.
Technical interview experience 3 - Deep Java + practical backend
Why Redis vs RDBMS for this read pattern? How would you refresh data daily?
Failure handling if download URL changes or zip is corrupt
Key takeaways: Older Zerodha assignments emphasise script + datastore + minimal UI - still a good practice template.
HR / culture interview experience 5 - Bootstrapped fit
Culture interview experience 5
Candidate profile: Final-round culture conversation; themes from InterviewEra and public CEO interviews.
Typical prompts
Why Zerodha instead of a VC-backed competitor?
Tell me about a project you owned from idea → deploy → iteration
How do you balance speed vs correctness when money or user trust is involved?
What do you know about Zerodha being bootstrapped, customer-first, and intentionally small?
What interviewers probe
Passion for the cause (helping Indians participate in markets responsibly) vs résumé padding
Comfort without a large support org - you debug, you document, you ship
Questions you ask them: team rituals, on-call expectations, how decisions get made without layers of PMs
Key takeaways: Generic “innovation” talk fails; specific product observations (Kite, Console, Varsity, tech blog posts) land better.
Sample answers for high-frequency questions
Use these as frames, not scripts.
Technical
Q. Simple order-book mental model. A. Bids and asks sorted by price-time priority; matching pairs compatible orders; emit trades and update holdings. Mention partial fills, cancellation, and idempotent order submission if retries exist.
Q. How do you show ownership in an interview? A. Narrate design → implementation → testing → deploy → post-launch fix with one metric (latency, error rate, users served). Say which decisions were yours vs the team’s.
Q. Rate limiter in one minute. A. Token bucket for bursts; fixed window is easier but unfair at boundaries; distributed setups use Redis with TTL; discuss 429 responses and client backoff.
Q. Stock price service sketch. A. Ingest ticks via Kafka (or similar), fan-out with pub-sub, WebSockets to clients, handle slow consumers with backpressure and reconnect state - you do not fan-out every tick through a relational DB.
Culture / HR
Q. Why Zerodha? A. Tie to correctness when moving money, learning on a lean team, and one concrete thing you read (engineering blog, product behaviour on Kite during market open). Avoid salary-first framing.
Q. What would you change if you had more time on the assignment? A. Pick one: integration tests, observability, input validation, deployment script - and say what you deliberately skipped for scope.
These links are research inputs we used to map assignment-first SWE loops - not copied page mirrors. We reorganized reports into process tables, outcome notes, and prep takeaways for Zerodha.
Ready to prepare? Ship one small fintech-adjacent service, defend it out loud, and read Zerodha’s primary sources - not recycled interview lists alone.
