About
AI Engineer at Intel Corporation building production LLM systems — retrieval, multi-agent workflows, and developer tools that engineers actually use. GSoC 2021 alumnus. I care about grounded retrieval, agentic harnesses, and evals that survive contact with a real engineering org.
Work Experience
Skills
Check out my latest work
Production LLM tools used by engineering teams, plus the open-source work I can actually link. Employer-internal systems don't have public demos — the write-ups stay high-level on purpose.
Multi-agent Debug Assistant
GPT-4 + AutoGen harness with RAG, root-cause reasoning, and image/text analysis. Helps engineers debug from messy production logs and docs instead of a clean demo corpus. Azure OpenAI backend.
Copilot + MCP Agent
GitHub Copilot agent plus a skills library. An agentic harness orchestrates MCP servers for ticket fetch, similarity search, and knowledge lookup so engineers get grounded context in the editor.
Regression Investigation Tool
Python GUI + CLI that correlates defect context with binary and code diffs to surface likely root-cause components. Shipped as a standalone desktop tool for engineers doing regression work.
CNN-LSTM Log Anomaly Detector
First-level log triage with a CNN-LSTM on an imbalanced dataset. Auto-generates reports with linked tickets. Dockerized Flask app behind NGINX with MongoDB.
A few things I'm proud of
Awards, selections, and talks — the work behind them is in the experience and projects above.
Innovation award
Intel Corporation
Recognized for shipping developer tools adopted across engineering teams, plus internal tech talks on the work.
Google Summer of Code 2021
FOSSology
Selected from ~8% of applicants to integrate ScanCode into FOSSology. Work merged into the main repository.