i build AI systems and products that are useful in the real world.

over the last few years, i've worked across healthcare, PR, jewellery, and edtech, shipping everything from LLM copilots and RAG systems to image-generation pipelines and agentic workflows.

• twain.ai, an AI jewellery design workspace that compresses months of design iteration into hours.
• Prompt Studio, a prompt analysis platform with a fully deterministic 7-dimension scoring engine.
• jev-studio, a CLI and MCP server for TypeSafe's Jev, listed on PyPI with 1K+ downloads.
• EasyEd, a student counselling chatbot that served 1K+ test users across 3 personas.
• Text Summarizer, PEGASUS fine-tuned for abstractive dialogue summarization.
• Nexsafar.ai, an AI travel companion submitted at Smart India Hackathon.

i'm an AI Engineer at M37Labs. along the way, i co-founded HealthyEnough, started devXsphere, mentored at GirlScript Summer of Code, and grew from member to president at DSDL, KIET.

1. 01

   i didn't get into AI because it was trending. i got into it because i was fascinated by the idea that software could learn from data and make decisions on its own.

2. 02

   the best way to learn is to build in public, share what works, and stay curious.← still true

3. 03

   outside of work, i mentor in open-source programs, contribute to developer communities, and help people break into AI. i also read a lot (books are half the reason i think the way i do) and travel whenever i find a reason to.

Agents

Claude Code, Codex, LangChain / LangGraph, and the OpenAI Agents SDK

Frontend

React, TypeScript, Next.js, and Tailwind

Backend

Python, FastAPI, and Flask

Models

PyTorch, TensorFlow, and HuggingFace

Data

PostgreSQL / pgvector, MongoDB, and Redis

Infra

AWS, GCP, Docker, Linux, and Git