AI Engineer & Full-Stack Developer
I don't follow tutorials. I train 768M-parameter LLMs, build stealth messaging apps, and create AI code assistants. Real engineering, no shortcuts.
Selected Work
Multilingual LLM built entirely from scratch. Custom tokenizer, training pipeline, and inference engine. No fine-tuning existing models—pure architecture engineering.
Secret messaging app disguised as a calculator. Hidden chats, encrypted storage, self-destructing messages. Invisible to anyone who doesn't know it's there.
AI-powered code assistant with autocomplete, error detection, and refactoring. Built with custom LLM integration for intelligent code understanding.
About
I'm an AI Engineer based in India. I build systems from the ground up—no templates, no shortcuts. My flagship project LAXMANA is a 768M parameter multilingual LLM trained entirely from scratch.
Currently open for new projects. I work directly with clients—no middlemen, no agencies. Just real engineering, real results.
PyTorch, TensorFlow, Transformers, CUDA, LLMs
Python, Node.js, FastAPI, PostgreSQL, Redis
React, JavaScript, HTML/CSS
React Native, Android
Contact
Have a project in mind? I typically respond within 24 hours.