Notes and articles on AI, engineering, and building dependable products.
A practical breakdown of preprocessing, evaluation, and deployment decisions that make ML systems stable in production.
Trade-offs between accuracy, latency, and power while shipping computer vision features on constrained hardware.
A framework I use to move from idea to production in small, measurable milestones without overengineering.