Senior Software Engineer focused on AI backend and ML platform work.
I build generative AI systems for subsurface geological modeling in Python, working with asyncio, multiprocessing, pandas, NumPy and SciPy, and multi-agent LLM systems built with LangChain and LangGraph. Before that, I was an exploration manager and chief geo-data scientist in oil and gas, which is where the older geoscience code in this account comes from.
Most of my professional work is confidential, so the public projects here are where I show how I build.
My1040NR: free, open source tax preparation for international students (Forms 8843 and 1040-NR, plus state returns). Everything runs in the browser on the user's device: documents are never sent to a server and never stored. No accounts, always free.
- numa.quest: turns a full sermon recording into captioned short videos, newsletter drafts, blog posts and small-group discussion guides.
- ML platform portfolio projects for my next role, published here as they land.
- Geoscience: oil and gas software I developed in industry, including SEG-Y and LAS readers and seismic tools. Mostly Delphi, some C#.
- Data-Science: machine learning and decline curve analysis for well data, including Niobrara formation depth prediction from very limited data at NeuDax.
Older learning projects (Udacity Android coursework from 2018 and university work from 2010) are archived, still readable if you want them.


