Building production AI systems that don't break in the real world — RAG pipelines, multi-agent backends, and full-stack applications shipped to production.
Three intelligent modules: automated flight disruption resolution via NLP, local experience discovery over OSM + Reddit + blog data into Qdrant, and multi-destination trip planning with dynamic re-planning and multi-user collaboration.
Real-time telemetry and analytics platform with live data ingestion, ML inference pipelines, WebSocket broadcast.
Hybrid FAISS + BM25 retrieval with 3-tier grounding system (strict/loose/none). Multi-user concurrent sessions, OAuth, semantic caching. Live deployed.
Personal productivity app that connects weekly task planning, daily life journaling, and focus tracking to generate an AI-graded weekly performance report. Built with React Native, Supabase, and the Anthropic Claude API — with server-side AI calls via Edge Functions, offline-first architecture, and Row Level Security for multi-user safety.
Turns any public LeetCode profile into an F1-style driver card rated out of 99, blending four scouting metrics (experience, racecraft, awareness, pace) with driver-archetype matching. Embeddable cards re-scout themselves on every load via Cloudflare Workers.
ML research & experimentation — F1 Race Predictor and a published recurrent-network paper.
View more →I build across the full stack with a focus on AI engineering — production RAG systems, multi-agent pipelines, and the infrastructure that keeps them running reliably at scale.
The goal is always the same: systems that work in production, not just in notebooks.
Final-year AI & Data Science engineer at Thadomal Shahani Engineering College, Mumbai. Hall of Fame Awardee. Published researcher at ICSCDS 2025.
I believe the most valuable engineers in the AI era are those who can own a system end-to-end — design the architecture, debug across layers of abstraction, and use AI tools as force multipliers without becoming dependent on them.
Currently building the F1 Real-Time Strategy Engine as a deep dive into Kubernetes, CI/CD, observability, and cross-platform frontends.
Let's connect
AI engineering and software engineering roles in Mumbai, remote US/UK, or Dubai.