Headline

Building multi-agent AI systems | React/Node/Express | Targeting Full-Stack & AI Engineering roles | B.Tech AI&ML, NIAT

About

I built a 5-agent AI system that takes a business owner's rough idea and turns it into a working pitch, deck, and WhatsApp-ready proposal — in one pipeline. That's the kind of problem I like: messy, real, and solvable with the right architecture. I'm a second-year B.Tech AI & ML student who spends more time shipping than studying for the sake of it. My recent work: BizPilot AI, a five-agent SME copilot (Intake → Context → Generation → Approval → Review) built for the TakeOver'26 hackathon, deployed live on Vercel and Render, using the Gemini API end to end. Before that: AutoFix AI, a six-agent diagnostic pipeline, and AI Research Crew, a sequential multi-agent system wired into MCP for tool use. Each one taught me something the last one didn't — mostly about where multi-agent systems break in practice, not in theory. Outside hackathons, I freelance: I've pitched and built dashboard and ERP-style tools for small businesses, including a school management system concept for a Bengaluru institution, and I'm extending an open-source holographic AI assistant UI into a full voice-and-chat product with real device control. I work across the stack — React front ends, Node/Express APIs, Supabase for data, LLM APIs (Gemini, Claude) for the intelligence layer — and I'd rather learn a concept by building it than by reading about it a third time. I'm looking for full-stack or AI engineering roles — internships or entry-level — where I can work on products people actually use, not just prototypes that die after a demo day. If you're hiring for that, or if you want to see BizPilot AI live, message me — I'll walk you through the architecture.

Location

Greater Bengaluru Area

AI stack

antigravity

codex

manus

midjourney

n8n

replit

Education

NxtWave of Innovation in Advanced Technologies