Naman Gupta
Backend engineer at Vegapay. Building credit infrastructure, distributed pipelines, and the occasional quant research project.
Based in Bangalore. Day job: credit card management systems — the data pipelines, reporting layers, and API surfaces that move information between core transaction systems and downstream services. Before Vegapay: BITS Pilani (ECE + Economics), competitive programming, and a stint at Université Laval on an international research fellowship.
Backend infrastructure
Java 21 + Spring Boot on a multi-tenant credit platform at 10K+ TPS and 99.99% availability. CCMS, reporting service, SFTP→Kafka→S3 ingestion.
Distributed pipelines
Kafka, Chronos, Postgres, S3. Built the data path that moves regulated credit data between core and downstream.
Quant systems
35+ predictive signals across US & China equities. Sharpe 1.25–2.91, Top 0.5% globally on AlphaVerse.
What I'm on this month
- Shipping
Extending the Vegapay reporting service to handle a second client onboarding — new file schemas, new SLAs, same pipeline.
- Reading
Designing Data-Intensive Applications (Kleppmann), and revisiting Marc Brooker's posts on queues and back-pressure.
- Open to
Senior backend / infra IC conversations at startups where the team is small enough that ownership is real. Remote (US/EU) or Bangalore.
Ask the site anything.
Three things worth talking about
Systems and research I've owned end-to-end — outcomes first, with the stack and shape of each. Full case-study writeups are in progress.
Vegapay Credit Platform — CCMS, Reporting & Ingestion
Backend on a multi-tenant credit platform running 10K+ TPS and 100M+ transactions a month at 99.99% availability. Own the Credit Card Management System and the REST/SOA layers that standardize responses across core transaction processors and third-party backends. Architected a reporting service end-to-end — system design, data modeling, implementation — plus the pipeline that ingests client CSVs over SFTP through Chronos and Kafka into Postgres and S3. Kafka reconciliation workflows hold consistency under concurrent load via message ordering and account-level locking. Drove initial request latency down ~80% to 30ms.
Cross-Market Alphas
Built and validated 35+ predictive signals on US and China equity markets. Discovered alphas with Sharpe ratios from 1.25 to 2.91, turnover up to 102%, cross-correlation under 0.7. Top 0.5% globally (rank 26) in AlphaVerse Premier League and global rank 302 in International Quant Championship.
NER for Legal Document Privacy
Published research on Named Entity Recognition-based privacy protection for legal documents. Built a custom NER pipeline using SpaCy, NLTK, and regex to automate sensitive entity detection across legal contracts and case files. Published in IJSRCSEIT, December 2023.
Four systems you can break in your browser
Build-your-own-X projects, each chosen to corner a different hard problem — concurrency, consensus, real-time convergence, AI infrastructure — and each running live right now. Open one and try to break it.
FrontRow
A seat-booking system that can’t oversell — even when a thousand people grab the last seat at once.
Raft lab
A real Raft consensus cluster you can kill, partition, and watch heal — live, in the browser.
Liveboard
A multiplayer canvas where every screen converges to the same picture, with no central lock.
Five-plus years of contests
Algorithmic problem solving has been a parallel discipline since school. Ratings below refresh live via the backend, which polls contest platforms on a schedule and caches the results.
- 2023ICPC Qualifier — overall rank 205, university rank 2
- 2022Google KickStart — global rank 1024
- 2023Meta Hacker Cup — global rank 1984
- 2023AlphaVerse Premier League (AlphaGrep) — university topper, global rank 26 (top 0.5%)
- 2023WorldQuant International Quant Championship — global rank 302 (Stage 2)
Notes, essays, and post-mortems
Long-form writing on the systems I've built — the design decisions, the trade-offs I weighed, and what I'd do differently.
Four systems, and what each one proves
Why I built four build-your-own-X backends instead of CRUD clones — and the one hard thing each one is meant to prove.
Zero oversold seats under 1,000 concurrent bookings
A two-layer no-double-book design — a Redis hold in front of a Postgres conditional write — plus a Kafka outbox, and a stampede that proves it never oversells.
A Raft KV store you can kill, partition, and watch recover
A from-scratch Raft implementation — leader election, log replication, the §5.4 safety rules — running five nodes over a transport you can break in the browser.
A real-time multiplayer canvas: a grow-only CRDT and Lamport ordering
Conflict-free convergence across clients and server instances — a grow-only-set CRDT, Lamport ordering, Redis fan-out — with 701 real messages relayed.
An OpenAI-compatible LLM gateway: routing, caching, and a circuit breaker
A drop-in multi-provider proxy that keeps serving when a provider falls over — model routing, response cache, SSE passthrough, and a per-provider circuit breaker.
Why we built our own report ingestion instead of using a managed ETL
The case for Kafka + Chronos over a SaaS pipeline when you’re moving regulated financial data — and where I’d revisit that call.
35 alphas later: what generalized, what didn’t
An honest review of which signal classes survived out-of-sample testing across regions, and the structural reasons most retail-discovered alphas fail.
From paper to pipeline: deploying NER for legal privacy at scale
Translating a research-grade NER model into something a law firm could actually run on a Monday morning.
Where I've worked
Vegapay Technology
Software DeveloperBackend engineering on a multi-tenant credit platform — 10K+ TPS, 100M+ transactions a month, 99.99% availability. Own the Credit Card Management System and its REST/SOA layers. Architected and shipped a scalable reporting microservice from scratch, built the SFTP → Kafka → S3 ingestion pipeline, and implemented Kafka reconciliation workflows using message ordering and account-level locking. Stood up a QA automation framework that cut regression time by 95% and enabled fully automated production releases.
MASTH (UltraHive Healthcare)
Chatbot / ML InternBuilt CBT-driven conversational modules in DialogFlow for a mental wellness chatbot serving ~1K users. Shipped a Flutter front-end with personalized activity tracking into the production Play Store app.
goGlocal
Backend Developer InternDesigned a Postgres schema handling 1.5M+ records with 70% test coverage. Shipped REST APIs sustaining 20K+ daily requests at 150ms average latency for an e-commerce marketplace.
Algoix Technologies
Full-Stack Developer InternBuilt a recruiter–learner–mentor platform from scratch in Django. 20+ data access endpoints serving 5K+ users.
Education and distinctions
Birla Institute of Technology and Science, Pilani
2020 — 2024B.E. Electronics & Communication Engineering and M.Sc. Economics. Founded the Alphas research group; selected for MITACS Globalink research at Université Laval, Canada.
- 2023MITACS Globalink Research Fellowship — Université Laval, Canada
- 2023Research publication: NER-based Privacy Protection (IJSRCSEIT)
- 2025Vegapay company-wide hackathon — Runner-up
Let's talk.
Most interested in backend or infrastructure-leaning roles at startups where the engineering bar is high and the team is small enough that ownership is real. Open to remote (US/EU) and on-site in Bangalore.
Book 15 min — pick any open slot