I build AI systems
that ship different
GenAI, agentic systems & ML, built end-to-end and shipped, not just notebooked.
View ProjectsSynthSales is an agentic B2B sales CRM that converts uploaded company CSVs into researched, scored, and contactable prospects through an 8-agent AI pipeline. Its FastAPI backend handles enrichment, email verification, outreach, reply tracking, meeting scheduling, access gating, kill-switches, and pipeline undo.
VOCA helps users save words, practice with spaced repetition, and track vocabulary growth. It includes a LangChain AI assistant with chat memory and one-click word saving.
Self-healing Airflow pipeline that uses a local LLM to detect and repair schema drift. Includes dbt transformations, PostgreSQL audit logs, and a Streamlit monitoring dashboard.
A Chrome extension for real-time chat over any YouTube video, built on a RAG pipeline with LangChain and ChromaDB. Semantic search runs over transcripts with adaptive context scaling and persistent chat history.
I'm a computer science student at Thapar who builds AI that does real work instead of living in notebooks: agentic systems, RAG pipelines, and LLM apps wired into production workflows.
My favorite projects are the ones where the AI acts, not just answers: a data pipeline that heals its own failures, a sales platform that runs its own outreach.
I've completed Stanford and DeepLearning.AI machine learning coursework, and I'm looking for ML and GenAI engineering roles where shipping matters. If you're building something ambitious with LLMs, let's talk.
B.E. Computer Science, Thapar Institute of Engineering and Technology, 2023-2027, CGPA 8.71
Have a role, a project, or just want to say hi? Drop a message and it lands straight in my inbox.