is research was primarily methodological and focuses on the interpretation of free-living waveform acceleration data. His projects included using deep learning to infer activity energy expenditure from wrist acceleration signals, and exploring how the resulting models can be repurposed for activity recognition.