Expect problems like detecting duplicate charges within a time window, building an O(1) in-memory cache with TTL and eviction, and maximising redeemable reward points under constraints. Interviewers want correct, readable code, a sensible data-structure choice, and a clear complexity analysis, plus the judgement to model a fuzzy product constraint precisely. CRED's engineering brand is polish, so sloppy code or hand-waving on edge cases stands out for the wrong reasons.
Recruitment process
- Recruiter screen or online assessment
- Technical coding round one on DSA
- Technical coding round two plus low-level design
- Hiring-manager round, then offer
Interview rounds
- Round 1 (60 min): coding round on DSA with a product framing
- Round 2 (60 min): harder coding plus low-level design
- Round 3 (45-60 min): design and behavioral round on quality and ownership
Clean and product aware framework
- Solve the DSA cleanly with the right data structure and stated complexity, and model the product constraint (payments, rewards) precisely
Practice questions
- 01Practise
Given a list of credit-card transactions, detect duplicate charges within a short time window. Optimise it and state your complexity.
StandardCRED technical rounds favour clean, correct code with a product-quality bar. Expect DSA framed around payments, rewards, and card data. - 02Practise
Design an in-memory cache with expiry (TTL) and an eviction policy. Which data structures give O(1) get and put?
Standard - 03Practise
Given reward points earned across many transactions, compute the maximum points redeemable under a set of constraints. Explain your approach.
IntenseCRED interviewers pair DSA with a product framing. Strong candidates model the constraints cleanly and reason about the greedy-vs-DP trade-off. - 04Practise
Design a pipeline that computes per-user creditworthiness features from bureau data, transaction history, and app behaviour, ensuring feature reproducibility for model training.
IntenseCRED data-engineer rounds probe feature pipelines with point-in-time correctness and reproducibility for credit ML models. - 05Practise
Write SQL to compute each user's on-time payment ratio over the trailing 6 statements, handling users with fewer than 6 statements gracefully.
StandardCRED tests window-function SQL over statement history with careful handling of partial windows. - 06Practise
How would you design a data pipeline that respects strict PII and financial-data governance while still enabling analytics and model training?
StandardCRED probes governance-aware data engineering: tokenisation, access controls, and lineage for sensitive financial data.
