I build AI-native backend systems

An AI agent that poses as a confused elderly retiree, engages real scammers in live conversation, and extracts UPI IDs, bank accounts, and phishing links from their own messages — turning reactive defense into active intelligence gathering.

Records in-person meetings, transcribes them live, and turns them into assigned, emailed tasks — without anyone taking notes.

I built an offline P2P payment system that encodes transaction data as audio tones between phones (MFSK modulation) — no internet required — with a hybrid UDP fallback and an offline-first Room sync queue.

A custom Gymnasium environment wrapping SUMO that benchmarks a PPO agent against fixed-time and queue-based heuristic controllers across 27 controlled runs. The heuristic won every metric — and that honest negative result is the actual finding, not a failure.

Procedurally generated cloud security audit environment that tests LLM agents on finding and fixing real vulnerability patterns — built for the OpenENV MetaHackathon.

Prevents double-spending when two payout requests hit the same merchant balance at the same instant — verified with real threading tests, not mocks.

Local-first inbox automation agent — reads Gmail via OAuth2, semantically classifies every email into custom labels using an LLM, applies them automatically, and runs on a schedule with zero manual input. Skips already-labeled threads to avoid redundant processing.

I built an API that takes call-center audio, transcribes it, scores SOP compliance, and flags payment-intent signals — designed specifically to avoid the numeric hallucination problem that plagues pure LLM scoring.

Emergency home services platform for Indian housing societies — three clients (Android, admin dashboard, API) sharing one REST API with role-based middleware, real-time tracking, and Razorpay payments.

I built a carpooling and P2P vehicle rental platform with real-time driver dispatch — Redis GEO proximity search, distributed locking to prevent double-dispatch, a NATS event bus, and full AWS infrastructure-as-code across native Android, Flutter, and a TypeScript backend.

I cleaned and analyzed 4.4M Aadhaar enrollment records — removed 602K duplicates, standardized 65 state-name variants down to 38 — and served the results through a 14-endpoint Flask API.

I built a real-time marketplace connecting local vendors across language barriers, with a provider-chain translation fallback (Google → Microsoft → AWS → mock) and Redis-backed WebSocket negotiation.

I built a bulk party invitation generator — a CSV guest list goes in, personalized cards come out, delivered via email and WhatsApp through a Celery queue.

I built an AI cybersecurity companion for students — it scans suspicious links/emails and teaches through a teach-back quiz.

I built a local tool that reads my Gmail, categorizes emails by sender with a local/cloud LLM, and auto-applies nested labels.

I built a bilingual (Hindi/English) career guidance Android app with a 4-service Node.js backend for Smart India Hackathon.

I built a tool where you upload a policy PDF, AI extracts the enforceable rules, and a rule engine scans live data for violations with human-in-the-loop review.

I built an Android companion app for IIT Bombay's Continuing Education office — browse courses, view certificates, find contacts.

I built a voice assistant that works in 13 Indian languages to help rural citizens find government welfare schemes.