Writing

This section will contain accessible technical essays, research case studies and research-notebook pieces. The aim is to explain not only successful results, but also the reasoning, uncertainty, failure modes and scientific decisions behind the work.

Case studies

• HyperFlux: Building a Scientific Digital Twin That Learns

Planned research notes

• When a scientifically plausible transfer model fails
• What does it mean for a machine-learning model to “learn” a physical parameter?
• What is a scientific computation worth?
• Detecting change when you cannot monitor everything