Why these 3
In 2026, Razorpay, PhonePe, and Swiggy collectively open more SDE roles in India than any FAANG satellite office. They pay at or above FAANG-India median for SDE-2 and above, and their interview loops are shorter (5 rounds vs 7 to 8). Getting through them depends on domain-specific prep, not generic DSA grind.
The big story of the last 3 years in Indian tech hiring is how much ground the domestic unicorns have closed on FAANG. Razorpay and PhonePe now pay L6-equivalent bands for SDE-3 roles. Swiggy and Flipkart staff engineers earn within 10 percent of their Google Bengaluru peers. The difference is that the interview loops test domain fit alongside algorithmic chops.
If you prep for these loops the same way you prep for FAANG (LeetCode top 150, Grokking System Design), you will pass the DSA rounds and fail the system design rounds. The domain-specific prompts (payment idempotency, dispatch geospatial indexing, wallet LLD) are where most candidates wash out. This guide covers what each round actually tests and how to prep round-by-round.
Razorpay: 5-round loop
Online assessment
HackerEarth or internal platform. 90 minutes, 2 DSA problems (medium to hard) plus 8 to 10 MCQs on OS, DBMS, networks. Scoring is algorithmic correctness plus time complexity. Target both passing test cases and an optimal Big-O.
Technical round 1 (DSA)
60 minutes live coding on CoderPad. One medium LeetCode-style problem (arrays, hashmaps, sliding window) plus follow-ups on complexity. Talk through brute force first, then optimise. Razorpay weights clarity of thought over one-shot solutions.
Technical round 2 (systems)
60 minutes. For SDE-2 and above: a payment-gateway-flavoured system design (idempotency, retry, exactly-once semantics). For SDE-1: deeper DSA plus OOP design (parking lot, rate limiter).
Managerial round
45 minutes. Project deep dive on your most complex past work. Expect questions on trade-offs, what would you do differently, and conflict handling. They probe for ownership signals.
HR and culture fit
30 minutes. Standard behavioural: why Razorpay, failure story, long-term goals. Razorpay has a strong internal culture around "customer obsession" and "hustle"; reference those values in concrete examples.
Prep priority: payment-domain system design, idempotency keys, webhook retry logic. Read the Razorpay engineering blog for at least 5 past posts before round 2.
PhonePe: 5-round loop
Online assessment
AMCAT or internal platform. 2 coding problems plus aptitude section. PhonePe includes a pseudo-code debugging round at fresher level that many candidates miss by focusing only on DSA prep.
Tech round 1
60 to 75 minutes. Heavy on data structures: trees, graphs, DP. PhonePe is known for asking at least one graph problem per candidate at SDE-2 and above. Review BFS, DFS, Dijkstra, topological sort.
Tech round 2 (LLD)
60 minutes of low-level design. Classic prompts: design a wallet, design a splitwise, design a payment reminder service. They want class diagrams, interfaces, and how you handle concurrency.
Tech round 3 (HLD)
SDE-2 and above only. Design a UPI transaction flow or a notification system at 100 million users per day. Focus on sharding, queue choice, and failure modes. PhonePe interviewers often ask about regulatory and compliance constraints.
Hiring manager
Behavioural plus culture. PhonePe emphasises "frugality" and "bias for action". Examples of shipping with constraints land well here.
Prep priority: graph algorithms (BFS, DFS, Dijkstra, topological sort), LLD patterns (factory, strategy, observer), HLD for UPI-scale workloads. Watch Arpit Bhayani on transaction isolation levels.
Swiggy: 5-round loop
Online assessment
HackerRank. 2 problems, 90 minutes. Medium difficulty. Swiggy weights correctness over edge cases at this stage; pass all visible test cases first.
DSA round 1
60 minutes live. One hard problem or two medium. Swiggy engineers are known for liking interactive problems: API design puzzles, stream processing, cache eviction. Prep beyond vanilla LeetCode.
DSA round 2
60 minutes. Deeper algorithmic. Past prompts: find the kth largest element in a stream, implement LFU cache, design a rate limiter for the order API.
System design
SDE-2 and above. Classic Swiggy prompt: design a food delivery dispatch system. Expect deep questions on geospatial indexing (geohash, quadtree), ETA prediction, and surge pricing.
Bar raiser and HR
Bar raiser is a senior from a different team. Mix of behavioural and one technical deep dive. HR round covers notice period, compensation, and location preference.
Prep priority: geospatial indexing (geohash, quadtree, R-tree), stream processing basics, cache eviction policies. Read the Swiggy Bytes blog on their dispatch system evolution.
Comparison: pay, difficulty, focus
| Company | Core strength tested | Pay bands | Loop difficulty |
|---|---|---|---|
| Razorpay | Payments domain depth, API design, idempotency | SDE-1: 18 to 28 LPA; SDE-2: 30 to 50 LPA; SDE-3: 50 to 80 LPA | 6.5/10 (LeetCode medium floor) |
| PhonePe | LLD and HLD both heavy, system design weighted | SDE-1: 20 to 32 LPA; SDE-2: 35 to 58 LPA; SDE-3: 60 to 95 LPA | 7.5/10 (graphs and design are make-or-break) |
| Swiggy | Real-world dispatch and streaming problems, less vanilla | SDE-1: 18 to 30 LPA; SDE-2: 32 to 52 LPA; SDE-3: 55 to 85 LPA | 7/10 (DSA plus domain creativity) |
| Flipkart | Classic big-tech prep, scale-first system design | SDE-1: 22 to 34 LPA; SDE-2: 36 to 60 LPA; SDE-3: 65 to 100 LPA | 7.5/10 (closest to FAANG bar in India) |
| Zomato | Product sense plus engineering, ML-adjacent problems | SDE-1: 18 to 28 LPA; SDE-2: 30 to 50 LPA; SDE-3: 52 to 80 LPA | 6.5/10 (behavioural rounds carry weight) |
Pay bands from Levels.fyi and AmbitionBox aggregated data, Q1 2026. Base plus stock, Bengaluru cost-of-living adjusted.
5 resume tactics that land interviews
Lead with payments or fintech keywords if you have them
UPI, NPCI, PCI-DSS, KYC, AML, idempotency, webhooks, reconciliation. Razorpay and PhonePe recruiters filter on these. If you built a payment integration at a previous role, surface it in the top 3 bullets.
Show production-scale numbers, not toy metrics
"Handled 12K RPS at peak" beats "built a scalable system". Indian unicorns assess engineering maturity through the scale you have already operated at. Add transactions-per-second, users, data volume wherever possible.
Name-drop the stack these companies actually use
Java/Spring (Razorpay, PhonePe, Flipkart), Go and Node (Swiggy, Zomato), Kafka, PostgreSQL, Redis, Cassandra. If you have production experience with their stack, call it out in the skills section and at least one bullet.
Open-source contributions count double
A single merged PR to Spring, Kafka, or a fintech library is worth more than 5 side projects. Link the PR directly in a "Selected contributions" section under Projects.
Keep the resume to 1 page for under 4 years, 2 pages after
Indian recruiters skim faster than US counterparts. Make the first 3 bullets of your most recent role carry the impact. Everything after the first page gets 10 seconds at most.
4-week prep plan
- Week 1 (DSA foundations): LeetCode top 75 focused list. Arrays, strings, hashmaps, 2-pointer, sliding window. 3 to 5 problems per day. Target: 80 percent solved without hints.
- Week 2 (graphs and trees): BFS, DFS, union-find, topological sort, segment trees, tries. Graph-heavy companies (PhonePe, Flipkart) test this weekly. 4 problems per day plus 1 hard on weekends.
- Week 3 (LLD): Pick 5 classic prompts (parking lot, ride-sharing, splitwise, wallet, rate limiter). Write class diagrams on paper. Practise verbalising trade-offs. 1 prompt per day, 45 minutes each.
- Week 4 (HLD plus mocks): Study 3 domain prompts (UPI flow, dispatch, food-delivery tracking). Do 2 mock interviews per week on Pramp or interviewing.io. Focus on narrating trade-offs out loud, not silence-and-code.
External references
Frequently asked questions
What is the total interview loop duration?+
How many LeetCode problems should I solve?+
Do these companies hire freshers directly?+
Is a referral necessary?+
What is the notice-period expectation?+
Which company is easiest to crack?+
Do they ask about projects or only DSA?+
Build a resume that lands the OA invite
ResumeBuildz surfaces the domain keywords Razorpay, PhonePe, and Swiggy recruiters filter on. Fintech-ready in 12 minutes.