CRED & Razorpay at a Glance
CRED
- Founded 2018 by Kunal Shah
- Valuation: ~$6B (2024 funding round)
- ~1,000 employees (lean, selective)
- Products: credit card rewards, CRED Pay, CRED Cash, CRED Travel
- Headquarters: Bengaluru
- Known for: extremely high bar, curated member base (750+ credit score)
Razorpay
- Founded 2014 by Harshil Mathur & Shashank Kumar
- Valuation: ~$7.5B
- ~3,500 employees
- Products: payment gateway, Razorpay Capital, RazorpayX, POS
- Headquarters: Bengaluru
- Processes 60%+ of Indian internet transactions
CRED Interview Process
Resume Screening / Referral
CRED doesn't hire from campus extensively — most hires come from referrals, LinkedIn outreach from their talent team, or direct applications with strong profiles (GitHub, published work, previous notable employers). Non-referral cold applications have a low response rate. Build genuine connections at CRED via engineering blog posts, meetups, or Twitter/X interactions.
Online Coding Assessment — 90 min
2-3 problems, Medium to Hard difficulty on LeetCode equivalents. Topics: graphs (BFS/DFS/shortest path), trees, dynamic programming, heaps, sliding window. CRED expects clean, well-commented code — not just a working solution. Brute-force approaches rarely score well; optimal time complexity is expected.
Technical Round 1 — DSA (60 min)
Deep walk-through of 1-2 problems. Interviewer challenges your approach: "Can you do better?", "What if constraints change?", "What's your space complexity?" Behavioral questions about your most technically challenging project are woven in.
Technical Round 2 — System Design (60 min)
Design a component of CRED's infrastructure: reward points engine, credit score lookup service, payment routing, or user segmentation. CRED values: scalability + data consistency + simplicity. Don't over-engineer. Explain tradeoffs clearly.
Culture / Bar Raiser Round (30 min)
Senior engineering leader assesses: depth of ownership in past work, intellectual curiosity, product thinking, values alignment with CRED's mission. Not a technical round — prepare honest, specific stories about impact.
Razorpay Interview Process
Online HackerEarth Assessment — 90 min
2-3 coding problems (Easy to Medium-Hard). Razorpay uses HackerEarth/CodeSignal. Focus: data structures, string manipulation, graph problems, time complexity optimization. Both test cases and edge cases are evaluated.
Technical Round 1 — DSA + Low-Level Design (60 min)
1-2 DSA problems followed by a low-level design (LLD) question: design a rate limiter, a notification system, or an idempotency key manager. Expects clean OOP code (class diagram + implementation sketch). LLD is a strong differentiator at Razorpay.
Technical Round 2 — System Design (60 min)
Design a payment gateway, webhook delivery system, or settlement engine. Razorpay specifically tests: idempotency, retry mechanisms, distributed transaction handling, and monitoring/observability for financial transactions.
Hiring Manager / Culture Round (30 min)
Team fit, leadership potential, past experience depth, why Razorpay. Kunal Shah's philosophy at Razorpay: "Solve for the user, everything else follows." Engineers who demonstrate user empathy and product ownership tend to do well.
DSA Preparation (Both Companies)
| Topic | Key Problems | Frequency |
|---|---|---|
| Graphs | BFS/DFS, Dijkstra, Topological sort, Number of islands, Course Schedule | Very High |
| Dynamic Programming | Coin change, Longest subsequence, Knapsack, Edit distance, Matrix chain | High |
| Trees | LCA, Serialize/Deserialize, Path sum, BST validation, Vertical order traversal | High |
| Heaps / Priority Queue | Kth largest, Merge K sorted lists, Median from data stream | High |
| Sliding Window / Two Pointers | Longest substring without repeating, Minimum window substring | Medium |
| Design Problems (LLD) | Rate limiter, Cache (LRU/LFU), Task scheduler, Parking lot, Library system | High (Razorpay) |
Fintech System Design Topics
Payment Gateway Architecture
Core challenge: exactly-once payment processing in the face of network failures, retries, and double charges.
- Idempotency keys for all payment APIs (client-generated UUID per payment attempt)
- Two-phase state machine: INITIATED → PROCESSING → CAPTURED | FAILED
- Distributed locking (Redis SETNX) to prevent concurrent processing of same payment
- Webhook delivery with exponential backoff and dead-letter queue
- Reconciliation service (async) to detect and fix inconsistencies between internal DB and bank systems
Fraud Detection System
Real-time scoring of transactions with low latency (<50ms) and high accuracy.
- Event-driven: transaction event → Kafka → real-time fraud scorer
- Rule engine (Drools / custom): velocity rules (10 txns/min from same card), merchant risk rules
- ML model serving: feature extraction (user history, merchant history, geolocation) → model inference → risk score
- Feature store (Redis for real-time features, Hive/BigQuery for batch features)
- Feedback loop: chargebacks → retrain model
Reward / Points Engine (CRED-specific)
Credit card spend → CRED coins → redeemable across merchants. Core challenge: consistency and auditability.
- Ledger-based model: append-only transaction log for all point credits/debits
- Idempotent credit: same card payment event must credit exactly once (dedup via transaction ID)
- Eventual consistency acceptable for display; strong consistency required for redemption
- Batch nightly reconciliation: bank statement parsing → coins credit jobs
- Expiry policy: background job marking expired points, with grace period
Rate Limiter for Payment API
Prevent API abuse and protect backend services from traffic spikes.
- Token Bucket (Razorpay favorite): smooth burst handling. Redis INCR + TTL for distributed counter
- Sliding Window Counter: more accurate, higher storage (per-second or per-minute buckets)
- Implementation: Nginx (lua-resty-limit-traffic), Redis (EVAL with Lua scripts for atomicity)
- Per-merchant limits, per-IP limits, global limits — tiered enforcement
- Response: HTTP 429 Too Many Requests + Retry-After header
Salary 2026
| Role | CRED (CTC) | Razorpay (CTC) |
|---|---|---|
| SDE-1 (0–2 years) | ₹30–50 LPA | ₹18–28 LPA |
| SDE-2 (2–5 years) | ₹50–80 LPA | ₹28–45 LPA |
| SDE-3 / Senior SDE (5–8 years) | ₹80–120 LPA | ₹45–75 LPA |
| Staff / Principal (8+ years) | ₹120–180 LPA | ₹75–120 LPA |
Frequently Asked Questions
