AI / ML Engineer · Full-Stack Developer
Crafting intelligent, full-stack experiences.
From concept to prototype.
I'm an AI/ML Engineering student architecting full-stack intelligent applications, asynchronous API backends, and predictive models.
I take projects from concept to working prototype with modern Python and JavaScript frameworks — Next.js and FastAPI on the web, TensorFlow and scikit-learn under the hood.
My focus is sharpening toward applied machine learning, local LLM inference, and enterprise-grade AI-safety workflows — the guardrails that make intelligent systems trustworthy.
Tools I reach for — from the model down to the metal.
Full-stack, AI-first products — from ETL engines to local-LLM tooling.
Multi-Agent Data Engine — Expo Winner 2026
An autonomous multi-agent data engine orchestrated with LangGraph, where agents self-heal through reflection loops and every generated step is audited in real time by an AST-based security pass. FAISS powers retrieval, Human-in-the-Loop guardrails keep the pipeline accountable, and a React Control Center makes the whole system observable.
Semantic ETL Engine — 2026
An enterprise-grade Semantic ETL engine with Human-in-the-Loop validation. Strict boolean guardrails and Pydantic schema enforcement stop LLM hallucinations, while self-healing LCEL loops recursively debug and correct AI-generated JSON. Framer Motion UI on Vercel, async FastAPI backend on Render.
Async Summarizer — 2026
A full-stack asynchronous AI text processor that summarizes massive document batches through local LLM inference (Phi-3). A fault-tolerant, sequential task queue in FastAPI manages flow and prevents out-of-memory GPU crashes under heavy load, behind a minimalist React UI with real-time progress polling.
FreeCodeCamp
IIT Hyderabad
Autonomous Multi-Agent Data Engine (M.A.D.E.) built with LangGraph, FastAPI, and FAISS. Features self-healing agent reflection, real-time AST security auditing, Human-in-the-Loop (HITL) guardrails, and a modern React Control Center.
Hackathon · Mahindra University
Always building the next one
This canvas is yours — scribble something. Pick a colour and draw. (It clears on refresh.)
Open to internships, collaborations, and interesting problems. My inbox is always open.
anmolyadav2906@gmail.com