full stack developer, exploring machine learning.

i build things for the web and write about the intersection of software engineering and ml.

i like building small, fast products for the web and working on the parts that actually make them useful: product logic, infrastructure, content systems, and practical ml features.

most of my work sits somewhere between software engineering and applied ml. i care more about useful systems, clean interfaces, and measurable outcomes than novelty for its own sake.

this site is intentionally minimal. i wanted the projects and writing to do most of the talking.

elevator-ai2026notion2vault2026emotion-recognition2025german-conversations2024markdown-to-html2024mira-auth2024purenotes2024shorts-automator2024
why i stopped fine-tuning and started promptingembeddings are all you need (for most search problems)running models locally changed how i think about mleval-driven developmentdata quality still beats model quality