Research

AI + Cancer Research

My research asks whether machine learning can find the patterns in cancer data that people cannot see by eye, and turn them into better decisions about treatment.

01Motivation

Liver cancer (hepatocellular carcinoma, or HCC) is among the leading causes of cancer death worldwide, and it often returns after treatment. I wanted to work on a disease where better tools could clearly change outcomes.

02The problem

Labs now collect huge molecular and imaging datasets for each tumor, but much of their structure is too complex to see by eye. Key steps, such as testing how a mutation changes a protein or which molecules might block a pathway, still depend on slow experiments and simulations.

03Approach

I use spatial statistics and clustering to map how immune cells and blood vessels are arranged inside tumors, graph neural networks to predict how mutations change protein–DNA binding, and generative models to propose small molecules that could block two recurrence pathways at once.

04Why it matters

If models can link a tumor's structure to a patient's outcome, and suggest which molecules are worth testing, researchers can spend lab time on the most promising ideas. The long-term goal is care that is earlier, more precise and more personal.

Where my research is headed

  • Medical AIModels that clinicians can question, validate and trust at the bedside.
  • Protein modelingLearning how structure and mutation shape function, faster than simulation allows.
  • Computational biologySpatial and single-cell methods that read tissue as a system.
  • Neuroscience applicationsBringing the same data-driven tools to questions about the brain.
  • Diagnostic systemsEarlier, more accessible detection built from data hospitals already collect.

Talks

Presenting the research

Recorded talks where I explain my cancer research to an audience of engineers and AI practitioners.

Interested in research, collaboration or speaking? Get in touch.

Next: Projects →