Six years of learning to ship.
From mechanical engineering at IIT to backend systems at Cars24. The roles, the projects, the things that broke, and what each taught me.
Designing payment orchestration microservices and C2C platform features. Java + Spring Boot + heavy event-driven architecture, with payment gateways and seller/buyer workflows that need to stay reliable at scale.
At Cars24 I work on two systems: a payment orchestration layer that smart-routes transactions across Razorpay, Juspay, Cashfree, and Paytm based on success rate, cost, and failover — and a C2C platform handling seller/buyer marketplace flows at thousands of users per minute. Both demand correctness, not just availability.
- Payment orchestration microservices — multi-gateway routing (Razorpay, Juspay, Cashfree, Paytm) handling 1k peak TPS at 99.9%+ reliability.
- Sales panel API optimization — dropped p99 latency from 1–2s to ~100ms; added bulk upload/download across multiple formats to streamline agent workflows.
- C2C platform — seller/buyer monetization, chat payments, lead generation, seller journey revamp. Supports thousands of users/min and 1000+ daily payments.
- AI chatbot integration — trained and integrated on C2C-specific data via transactional APIs; enables payments, refunds, and order ops directly through chat.
- Verify-360 — Spring Boot + PostgreSQL document verification platform backed by Amazon S3; legal officials review/approve/reject vehicle documents by permission level.
- Platform reliability — centralized config module and logging across services; raised unit + integration test coverage from 0 to 50% on critical services. Integrated Clevertap (notifications), Spyne (image webhooks), and Acefone (call logs); parallelized C2C BFF-layer APIs in Node.js by user role.
Payments taught me that "eventually consistent" is a luxury you don't always get. Some systems demand correctness now, every time.
Built and owned test automation infrastructure for Sprinklr's Workforce Management module. Cut sanity run time from 12 hours to 2 hours across 1500 test cases. Engineered a JSON-driven API framework that eliminated DTO boilerplate entirely.
- Sanity pipeline optimization — cut 12-hour runs (1500 test cases) down to 2 hours through optimized Jenkins job orchestration.
- Spring Boot admin panel — company-wide test case management with alerts, recurring runs, and reports; handling 100s of users and 1000s of test runs.
- JSON-driven API automation framework — single JSON file executes full test suite across 7 flows (if/for/switch/payload/headers/query params) using strategy + factory builder patterns. Eliminated DTO boilerplate entirely.
- AI output validation — test modules for AI-generated forecasting and scheduling outputs; validates ~100 constraint combinations per input form across 10+ dropdowns.
Building test infrastructure is engineering, not "just QA". Bad tests kill velocity faster than bad code.
First full-time engineering role. Built HRMS modules with a team of 3, replacing an existing portal and saving ~$50K/year in licensing. Shipped tools used by 900+ employees.
- HRMS portal replacement — Payroll, Onboarding, Leave Management, and more with a team of 3; cut ~$50K/year in licensing costs.
- Expense & Access Management — automated tool access provisioning, used by 900+ employees.
- Analytics overhaul — integrated data from multiple databases and built a SQL query generator for visualization; dropped report generation from 1–2 hours to minutes.
First job is where you stop coding to learn and start coding to ship. Different skill.
First taste of production. Integrated analytics tools for a 3x interactivity uplift and automated GitLab → AWS EC2 deployments via Docker scripts, cutting deploy time by ~70% to ~5 minutes.
- Analytics stack — integrated Microsoft Clarity, Google Analytics, and Firebase to track user behavior; 3x interactivity uplift.
- Deployment automation — automated GitLab → AWS EC2 deploys via Docker scripts; cut deploy time by ~70% to ~5 minutes.
Watching real users break my code taught me more in 5 months than two years of college.
Studied Mechanical Engineering on paper, but spent more nights debugging code than CADding gearboxes. Department Rank 3 (CGPA 8.33). Designed the braking system for the BAJA SAE'21 ATV. Led a 35-person logistics team for Varchas '22.
The most useful course I "took" at IIT was the one I didn't enroll in — late nights, Stack Overflow, and the discovery that I liked computers more than gearboxes.
Six years in, four companies, one career pivot. The pattern, looking back, is simple: pick the next thing that scares you a little, then write down what you learned along the way.
The notebook continues.
Hiring for an SDE-2 backend role? Let's chat.
I'm building toward systems that scale, teams I respect, and problems that don't have textbook answers. If your role fits any of that, I'd love to hear from you.