I’m an AI & Software Engineer with a founder’s ownership mindset and a systems engineer’s discipline.
- I design assuming models fail
- I optimize for correctness, privacy, and constraints
- I build end-to-end systems, not toy demos
- I value depth over hype and shipping over slides
My engineering mindset is shaped by research-grade environments (DRDO) and by building systems meant to survive real-world usage, not just presentations.
- Research-first engineering culture
- Strong emphasis on validation, reliability, and assumptions
- Exposure to non-negotiable correctness constraints
- Bias toward failure-aware and security-conscious system design
This is where I learned to think before scaling.
- Languages: Python, C++, C, JavaScript, TypeScript
- Foundations: DSA, OOP, system design
- LLMs, RAG pipelines, FAISS
- Chunking & retrieval strategies
- Prompt & context engineering
- Hallucination mitigation (design-level)
- FastAPI, REST APIs
- PostgreSQL, MySQL, Redis
- Docker, Linux, Git
A privacy-preserving, offline medical document intelligence system.
Problem
- Most medical AI tools assume cloud access and data leakage
- Clinics need offline, auditable, secure solutions
What I Built
- OCR → chunking → embeddings → encrypted FAISS
- AES-256-GCM encryption with per-record IVs
- Decryption only in memory
- RAG pipeline using Med-R1 8B
- FastAPI backend
- Electron / React frontend
Scope
- Administrative & clerical workflows only
- No diagnosis (by design)
This project is built with production constraints, not hype metrics.
- Applied RAG quality optimization
- Local-first AI for regulated environments
- Secure system boundaries
- Failure-aware architectures
- Systems that remain reliable under pressure
- 🏆 Smart India Hackathon — Back-to-Back Winner
- 🧪 DRDO Internship
- 📚 Strong grounding in systems, DSA, backend architecture
- 🧠 Comfortable discussing trade-offs, limits, and failure modes
Reach out if you’re:
- Building serious AI systems
- Operating in security- or privacy-critical domains
- A US startup, research lab, or deep-tech team
- Looking for someone who owns outcomes end-to-end
