Autopulse
Conversational copilot for automotive diagnostics — a RAG chatbot that answers questions about OBD2 and vehicle data. KPIT competition submission.
View Project →Portfolio & Résumé — Final Year CSE · Machine Learning · Full-Stack
I'm Muhammed Shamil — a final-year computer science student who turns curiosity into working software: RAG systems, machine learning projects, payment platforms and an LLM inference engine. Right now I'm recreating ChatGPT from scratch.
Conversational copilot for automotive diagnostics — a RAG chatbot that answers questions about OBD2 and vehicle data. KPIT competition submission.
View Project →Centralised payment system for friend groups — tracks who owes whom and settles debts with the minimum number of transactions. Backend plus Flutter & web front-ends.
View Project →Symbolic regression research — developing a model that discovers interpretable mathematical expressions from data instead of black-box fits.
View Project →An LLM inference engine built from the ground up — built to understand transformers from the inside out rather than just call them.
View Project →A simple, humble blog interface that lives inside your Windows terminal. No browser, no fuss — just text and a blinking cursor.
View Project →"Thy end is now!!" A gloriously useless full-stack idea with a Flutter front-end. Proof that not every project needs a point.
View Project →A Python experiment in generative art — drawing the traditional Kerala pookkalam, infinitely.
View Project →Currently Building
Recreating ChatGPT from scratch —
fully pretrained & finetune-ready for public use.
I'm Shamil, a final-year Computer Science & Engineering student from Thrissur, Kerala — the kind of person who reads a paper and then rebuilds it from scratch to make sure it's real.
My interests are deliberately wide: transformers and machine learning today, game development with Lua tomorrow, and a healthy amount of full-stack engineering in between. Ask me about software, AI — anything interesting. What I don't know, I learn.
B.Tech CSE · 2022 — 2026 · Thrissur, Kerala, India
Designing and training models — from symbolic regression to transformer architectures.
Building, fine-tuning and running LLMs — including an inference engine from the ground up.
Retrieval-augmented generation over domain data, with LangChain and vector stores.
APIs and services in Python (FastAPI) and JavaScript (Node.js).
Interactive web apps with Next.js and cross-platform mobile with Flutter.
MySQL, vector stores and AWS-backed persistence, backups and deployment.
“The universe is asymmetrical and unbalanced.”
— A fun fact I keep coming back to