About
I work on research software, applied machine learning, and technical writing where the process matters as much as the final artifact.
My work sits between software systems and research practice. I am interested in tools that help technical work remain inspectable: research environments that preserve context across reading and analysis, model pipelines that make assumptions and evaluation visible, and written records that explain why decisions were made rather than only presenting the result.
The projects collected here are deliberately modest in tone. Refract is a research workspace built around durable sessions, source-linked evidence, and analysis surfaces. The Alzheimer MRI study is a notebook-first machine-learning project where the value is the visible sequence of data inspection, preprocessing, training, and evaluation. The writing exists to make those choices easier to examine later.
Working interests
- research workflows that connect source material, saved evidence, comparison, and statistical analysis
- applied machine learning in medical and scientific contexts, especially where evaluation and limitations need to stay visible
- interfaces that reduce the distance between a result and the evidence or data that produced it
- technical writing that explains tradeoffs without making the work sound more complete than it is