Startup · Bengaluru
Zerodha interviews are assignment-heavy and focus on ownership, backend engineering, and deep product thinking.
Process: Take-home Assignment → Technical × 2 → Culture Fit
What is Zerodha's engineering culture and what kind of engineers do they look for?
Tip: Zerodha: India's largest stock broker by active clients. Culture: extreme ownership, small teams building large things, no micromanagement. Engineers often own full products solo. Interview style: take-home assignment (real problem), deep technical discussion on your solution, then culture-fit conversation focused on ownership and simplicity.
How would you design Zerodha Kite's order management system to handle market open volatility (9:15 AM surge)?
Tip: Challenge: 10× normal order flow in 15 seconds at market open. Strategy: pre-validate orders before 9:15 (margin check, symbol validation) → queue in Redis/Kafka → dispatch to broker in burst mode. Rate limiting per user. Priority queues: limit orders pre-queued at 9:14 for instant dispatch at open.
What are the key differences between Zerodha's Go-based backend and the typical Java/Node.js stacks?
Tip: Go advantages at Zerodha: low-latency order processing, goroutines for concurrency (no thread overhead), small binary size, fast compilation. Zerodha's Kite platform is heavily Go-based. Knowing Go basics (goroutines, channels, defer) is a genuine differentiator for Zerodha interviews.
Explain Zerodha's approach to open-source. What projects have they contributed to?
Tip: Zerodha open-sources significant parts of their stack: Kite Connect API (broker API), Gokiteconnect (Go client), kite.trade (Python client), and multiple financial data tools. They also maintain Mandi (go-based task runner) and GoCraft. Mentioning their open-source work signals genuine research beyond their Wikipedia page.
Write code to implement a simple moving average (SMA) and exponential moving average (EMA).
Tip: SMA(n): rolling sum of last n prices / n — use a deque for O(1) window update. EMA: EMA = price × k + EMA_prev × (1-k) where k = 2/(n+1). EMA reacts faster to recent prices. These are core technical indicators — Zerodha's interview may ask you to implement charting algorithms.
How does margin calculation work in equity trading and how would you implement it?
Tip: SEBI mandates: equity delivery requires full cash (100% margin). Intraday: MIS product type, 5× leverage (20% margin). F&O: span margin + exposure margin from exchange. Zerodha's margin engine runs pre-trade validation: if available_margin < required_margin → reject order. Real-time margin update on every fill.
What is a WebSocket and how does Zerodha use it for real-time market data in Kite?
Tip: WebSocket: full-duplex communication over a single TCP connection. Zerodha Kite WebSocket: subscribe to symbols → receive binary-encoded tick data (price, volume, bid/ask) at 1 sec intervals. Binary encoding (not JSON) for bandwidth efficiency. Client reconnects with exponential backoff on disconnect.
Tell me about a project where you owned the entire lifecycle from design to deployment.
Tip: Zerodha values full ownership. Show you thought about: requirements gathering, architecture decisions, implementation, testing, deployment strategy, monitoring, and post-launch iteration. Even a college project counts if you can speak to all phases. The question probes for ownership mindset, not just coding skill.
How would you build a backtesting engine for trading strategies on historical data?
Tip: Core loop: iterate over historical OHLCV bars → feed to strategy → generate signals (buy/sell) → simulate execution (price slippage, brokerage) → update portfolio state. Data: compressed binary format (HDF5, Parquet) for fast I/O. Output: equity curve, Sharpe ratio, drawdown. Zerodha's Streak product is a no-code backtesting tool.
What is the difference between Postgres and Clickhouse? When would Zerodha use each?
Tip: Postgres: OLTP, row-oriented, ACID transactions — for user accounts, orders, portfolio state. ClickHouse: columnar OLAP, extremely fast aggregations on time-series — for tick data analytics, trading volume reports, user behaviour analysis. Zerodha uses both: Postgres for transactional core, ClickHouse for analytics dashboards.
Zerodha Broking Limited interviews follow a 3-round process. Here is what to expect and how to prepare for each stage.
Upload your resume and get questions scored across technical depth, communication, structure, confidence, and relevance — the same criteria Zerodha panels use.