About
I like building systems that have to actually work. The kind where someone is waiting on the other end of an API call, and a 200ms delay is the difference between a good day and a bug ticket. My path into software engineering wasn't the obvious one. I studied Electrical Engineering at IIT Dhanbad, but somewhere between a data structures assignment and my first real backend project, I realized I'd rather spend my evenings debugging a caching layer than solving circuit problems. So I leaned in. Competitive programming, side projects, reading other people's code on GitHub at 1am. By the time I graduated, I wasn't an EE who could code. I was a backend engineer who happened to have an EE degree. At ZS, I get to do the thing I find genuinely fun: making search and AI systems faster and smarter. I've shipped Java Spring Boot microservices powering hybrid semantic search across 5M+ biomedical records, redesigned a RAG pipeline into an agentic LangGraph architecture that cut latency by 40%, and built an MCP server that other teams now plug into for their own RAG work. On the side, I built QuillFlow, a production-grade agentic RAG system with hybrid retrieval, multi-tenant auth, and a two-tier semantic cache that cuts repeat query latency by 10x. It's the project I point to when someone asks how I actually think about building things end-to-end. What I work with day-to-day: Java, Python, Spring Boot, FastAPI, LangGraph, GraphQL, PostgreSQL, Redis, Elasticsearch, Qdrant, Docker, Kubernetes, GitLab CI/CD, and a healthy obsession with system design. I'm always up for a conversation about backend architecture, agentic systems, or how to make RAG pipelines stop hallucinating. If you're building something interesting in those spaces, or hiring for a team that is, my inbox is open.
Work experience
Business Technology Soultions Associate at ZS, Jun 2025 - Present · 1 yr 5 mos, Pune District, - Built a PoC uniting Google Agent Space, AWS Bedrock, and Databricks Agents to deliver enterprise Q&A and insights from structured data, advancing GenAI-powered content search for next‑gen AI initiatives. - Transformed a basic RAG pipeline into a robust LangGraph-powered retrieval system, leveraging AI agents and parallelized queries to cut response times by ~5 seconds and deliver concise, high-quality summaries.
