I BUILD AI SYSTEMS. END TO END.
Hey, I'm Kushal — a BTech CSE (AI & DS) student at MIT-WPU and a full-stack & ML engineer. I build AI agents, backend systems, automations and computer-vision pipelines that ship to production — from OCR pipelines to voice tutors to auto-scaling ML.
CS STUDENT & FULL-STACK AI ENGINEER
I'm a BTech CSE (AI & Data Science) student at MIT-WPU who builds end-to-end — from FastAPI backends and Postgres pipelines to ML models, computer vision, and LLM-powered agents that actually ship.
I've built an enterprise OCR + court-data ETL platform for bank due diligence, a voice AI tutor on a Neo4j knowledge graph, auto-scaling ML for Mitsubishi, and won 1st place at the BITS Pilani national hackathon among 100+ teams.
"I don't stop at a model — I ship the whole system around it."
PROJECTS THAT SHIPPED TO PROD
Property Due Diligence
Enterprise OCR + court-data ETL platform for bank loans. Async pipelines, SSO/JWT entitlements, WOPI editable reports — cut verification time by 35%.
Voice AI Tutor
Voice-interactive tutor with speech recognition, reasoning & whiteboard visualization on a Neo4j knowledge graph. Boosted learner engagement by 40%.
BITS Pilani Win
1st place among 100+ teams at the BITS Pilani national hackathon — production-ready end-to-end prototype delivered in 36 hours.
FROM PROBLEM TO PRODUCTION
Scope
I dig into the real problem and constraints — data, scale, latency — and design the architecture before writing a line.
Build
Backend, models, and frontend built together — typed APIs, tested pipelines, containerized and CI-ready.
Ship
Deployed on Docker/Kubernetes/AWS with monitoring & dashboards — so it runs reliably, not just on my laptop.
Scale
Measure, profile, and harden — auto-scaling, caching, and anomaly detection to keep it fast under real load.
LET'S BUILD SOMETHING.
Open to internships, freelance builds, and collaborations. Drop me a line — I reply fast.