Backend engineer at Nomura — Java, Spring Boot and Apache Camel in Global Markets IT (Prime Brokerage).
I build high-throughput, low-latency systems where a missed run is a regulatory problem, not just a bug.
I work in Global Markets IT (Prime Brokerage) at Nomura, on the systems behind regulatory reporting, client reporting, trade reconciliation and enterprise integration. The work is Java at the point where correctness and throughput both matter:
- ⚡ Regulatory trade persistence microservice handling 60M+ trade records — cut insertion latency from 3 s → 40 ms through batching, transaction management and direct JPA persistence.
- 🛡️ SME for 4+ business-critical Java and Apache Camel adapters across workflows — production support, incident triage, debugging, releases and deployments. The pager is mine.
- 🚪 Reusable integration gateway framework built on the Template Design Pattern: configurable routing, transformations, centralised exception handling, Grafana metrics and health checks. Benchmarked at 462 HTTP messages/sec and adopted across multiple downstream services.
- ☁️ AWS vendor-feed integration — Java and Camel adapters polling SQS over HTTP and pulling files straight from S3, with validation gates before anything reaches downstream systems.
- 🐍 Python automation framework over the Jira and GitLab APIs — automated dependency upgrades, multi-repository code changes, commits and Merge Requests across 20+ Spring Boot microservices.
I built Sonar AutoFix — an LLM-powered CI/CD remediation platform using Dockerized GitLab Runners, LLM workflows and Javalang parsers to read SonarQube findings, generate fixes and raise Merge Requests for developer review. It went into production, cleared ~40% of our SonarQube issues across enterprise microservices, and won a Nomura Star quarterly recognition award.
That is the shape of how I work with AI: pointed at a real engineering problem, shipped behind review, measured. The projects below are the same instinct, in public.
| Project | What it is |
|---|---|
| CodeRAG | A code-aware RAG system that models source code by structure — method signatures, call graphs, static analysis — instead of flat-text chunking, so retrieval follows real code relationships. Python · FastAPI · Sentence Transformers · ChromaDB · React · TypeScript |
| ai-engineering-playbook ⭐58 | A zero-to-100 learning path for applied AI engineering — RAG, embeddings, vector search, agents, MCP and the production engineering around them. 56 pages, built as a searchable site. |
| CraftMyCV | Resume generator that tailors a CV from a candidate profile plus a target job description. Java 17 · Spring Boot · React · Python AI service · Docker |
| ChoreoAI | Open-source production layer for LangGraph — budgets, guardrails, retries, evals and observability for multi-agent systems. Companion CLI: choreoai-cli. Early, building in public. |
| TCS-NQT-CODING-SHEET ⭐951 · Infosys-SP-Coding-Questions ⭐134 | Curated Java solution sets for the TCS NQT and Infosys SP hiring tests. ~1,100 stars combined — still the most-used thing I've put on GitHub. |
Languages
Backend
Databases
MS SQL Server in anger: views, stored procedures and query tuning.
DevOps & Tools
Frontend
AI / LLM engineering
Concepts — Distributed Systems · Event-Driven Processing · Concurrency & Multithreading · Batch Processing · Transaction Management · Caching Strategies · Performance Optimization · OOP
B.Tech, Electronics & Communication Engineering — VIT Vellore (Sep 2020 – May 2024), CGPA 8.71/10.
Before Nomura: Software Engineer Intern at Wizergos, where I re-engineered a customer-facing chatbot from AngularJS to React.js — Hooks, component-based architecture, proper state management — and improved response performance by 50% through frontend and API-integration work.
Led weekly coding sessions for the Tech Enthusiasts Club. 5+ skill badges from IBM, Cisco, Google and AWS.