Data-first vision models
Why I obsess over dataset quality before tuning architectures, with checklists for balance, leakage, and robustness.
Read noteResearch notebook
Notes on computer vision research, evaluation, and applied ML practice.
Technical essays from research and deployment work. I write to clarify ideas, document tradeoffs, and share repeatable workflows.
Topics
Computer Vision
Model Evaluation
ML Systems
Research Process
Want to collaborate or quote something? Drop a line and I will reply quickly.
Email meLatest posts
Practical notes on data quality, model reliability, and decision-making in applied research.
Making experiments legible
Playbooks for writing logs, tables, and visuals that let collaborators reproduce results without endless back-and-forth.
Read noteBridging academia and product
How to translate research findings into product questions, metrics, and milestones that teams can ship against.
Read notePhotography as model training
Lessons from shooting street and landscape photos that map surprisingly well to dataset design and labeling rigor.
Read noteJoin the conversation
I welcome technical discussion, peer feedback, and collaboration on computer vision research problems.