Asif Khan

Machine Learning & AI for Medicine Postdoctoral Fellow · Harvard Medical School

I am a Postdoctoral Fellow at Harvard Medical School, working with Chris Sander. Before Harvard, I completed my PhD with Amos Storkey in the Bayesian and Neural Systems Group at the University of Edinburgh. My doctoral thesis was on the geometry for deep representation learning.

My research lies at the intersection of machine learning and biomedicine, addressing two critical challenges in cancer biology: early detection using longitudinal patient records and optimization of combination therapies through mechanistic models of drug response. Late-stage diagnosis remains the primary cause of cancer mortality, as many patients present when curative treatment is no longer possible. By training AI on large-scale clinical data, our work aims to identify high-risk individuals who can benefit from earlier intervention and effective treatment strategies. My current work is centered around following topics:

• Representation learning from longitudinal EHRs: Developing foundation models that encode patient histories into continuous representation spaces of patient health states for downstream survival and risk assessments.
• Uncertainty-aware cancer risk prediction: Building well-calibrated, robust, and generalizable models that capture distribution shifts across hospitals and populations.
• AI-guided therapy design: Learning differential equation models that capture molecular dynamics under drug perturbations to inform combination therapies.