Software engineer · San Francisco
I build AI for places where a wrong answer is expensive.
I’m Jeevan. At TaxGPT I build the retrieval behind tax agents used by tens of thousands of tax professionals. Before that I founded a startup, fine-tuned a legal model for US insurance regulators, and studied computer science and economics in seven countries.

1st of 156 AWS MCP/A2A Hackathon1st of 100+ NEAR AI Useful Agents HackathonSSRN working paper on AI and junior developers3.94 GPA Minerva University, CS and economics$5,000 NAIC Foundation Scholarship7 countries lived and worked in
Right now
Teaching agents to cite state tax law.
Each US state publishes its tax code its own way. I crawl, chunk and embed it so an agent can answer a question and point at the statute. Read the case study
134
pull requests merged in my first 13 weeks
90%+
faster crawling per search, after a rewrite
7
tools in an MCP server approved for the Claude directory
134
pull requests merged in my first 13 weeks
Selected work
All work & experience →Four things I’d show you first.
The path here
Nine stops, seven countries.
- KathmanduOrigin
- San Francisco2022 – 23
- SeoulFall 2023
- BerlinSpring 2024
- Kansas CitySummers 2024 & 25
- Buenos Aires2024 – 25
- TaipeiFall 2025
- HyderabadSpring 2026
- San Francisco2026 – Now
Take a break
Run across San Francisco’s rooftops.
Fog City Rush is a browser game I built from real OpenStreetMap data: parkour, ziplines, cars and a seaplane under the bridge. It plays right here.
Play it →Writing
All writing →Recent writing.
Apr 1, 2026Digital Bridges: Overcoming Nepal's Structural and Political Barriers to Economic MobilityWhy Nepal exports its workers, and five digital tools I built to lower the costs that geography and politics impose.Dec 16, 2025The Complexity of Discontent: Analyzing Nepal's Youth-Led Political UnrestA complexity-science reading of Nepal's Gen Z protests, where cheap digital coordination met a closed political cartel.Apr 25, 2025Increasing and Reforming Foreign Aid for Sustainable Global DevelopmentAn argument that rich countries should give more aid, not less, and spend it on human capital, institutions and local ownership.