Engineer, product builder, systems thinker.
Background
I'm an AI Software Engineer in San Francisco, currently building production AI products and internal agent infrastructure at Headline. My recent work includes launching a deck publishing platform, shipping an eight-provider email-finder waterfall that raised outreach match rates to 95%, and completing a zero-downtime Elasticsearch reindex. I previously interned at Headline, where I built a conversational market map with eight Mastra agents and a sandboxed file I/O layer for AI agents.
Before Headline, I was the Founding AI Engineer at VOCA Health, where we placed 5th overall and 2nd in the Health category at the Cozad New Venture Challenge 2025, securing $20,000 in seed funding for an iOS app that detects vocal disorders for 500+ beta users with 89% diagnostic accuracy. I began my career at Deloitte, building 65+ REST APIs and full-stack features in .NET.
Education
- Master of Science in Information Management (Computer Science), University of Illinois Urbana-Champaign — May 2026, GPA: 3.7/4.0. Coursework: Database Administration & Scaling, Operating Systems, Natural Language Processing, Generative AI.
- Bachelor of Engineering in Electronics and Communication Engineering, Sri Sivasubramaniya Nadar College of Engineering, Anna University — May 2022, GPA: 8.526/10.0. Coursework: OOP, Data Structures & Algorithms, Computer Architecture, AI/ML, Web Development.
Skills
- Languages: C, C++, C#, Go, Java, JavaScript, Python, Rust, TypeScript
- Web development: HTML, CSS, .NET Core, AngularJS, React, Node.js, NestJS, Next.js
- Cloud: AWS, Azure, Docker, GitLab, Firebase, Kubernetes, Render
- Databases: Microsoft SQL Server, MySQL Workbench, PostgreSQL, Neon, MongoDB
- Machine learning: TensorFlow, PyTorch, NLP, Keras, scikit-learn, PySpark, OpenCV, NumPy, Pandas
- AI development: Claude Code, Cursor, Codex, Grok Bot, Devin, MCP
How I work
- Ship end-to-end — I like owning products from ingest and storage through hosting, export, and the final user workflow.
- Engineer past the happy path — zero-downtime reindexing, backfill recovery, and isolated agent sandboxes matter as much as the demo.
- Measure the impact — 95% outreach match rate, 89% diagnostic accuracy, and 50 ms shaved off API response times.