Joining

Join the group!

• PhD Applicants: If you are interested in doing a PhD with me, feel free to reach out to me. There are a variety of routes to doing a PhD:
  • StatML Center for Doctoral Training
  • Math Department
  • There are also other centers for doctoral training on AI4Health and Collaborative Computational Modelling at the Interface, which I may be able to admit students from.
• Imperial Msc Students: If you are interested in doing a masters thesis project with me, reach out. I am advising masters students in multiple programs including I-X, statistics, and computing.
• Imperial Undergrads: If you are an Imperial undergrad interested in working with me, send me an email with your resume, transcript, links to any projects you’ve worked on, and a paragraph about your interests. I also recommend looking into the UROP program.
• Postdocs: Postdoc funding opportunities are currently more limited, so I would encourage looking into Schmidt AI in Science Fellowsips, as well as external fellowships.

Here is the learning agreement I use with students.

Possible project directions

I am interested in designing methods for real-world decision-making problems. I’m particularly interested in decision-making in environments that are stochastic and lack accurate mechanistic models of the environment (like healthcare, recommendation systems, social sciences—in contrast to video games and robotics). Some directions I’m currently particularly interested in:

• Using foundation models and pretraining techniques on historical data for more efficient exploration, decision-making, and policy improvement. See this paper for our initial efforts in this direction. I’m particularly interested in exploring if these methods can be used in small data settings.
• Collaborating with organizations interested in using decision-making algorithms. I recently started working with American Express and neonatal care physicians.