TL;DR — I am a scientific innovator at the intersection of AI and protein design, developing novel biomolecules for applications in precision medicine and environmental sustainability.

Present

I lead the Generate group in the Marks Lab at Harvard Medical School. Our research combines generative AI, computational biochemistry, and protein engineering to solve complex biological challenges. My recent work has focused on multimodal and retrieval-augmented protein language models, transfer learning across fitness landscapes, and enzyme design.

Not so long ago

I completed my PhD in the Oxford Applied and Theoretical Machine Learning Group, under the supervision of Yarin Gal. My doctoral research evolved from general machine learning to developing AI models addressing questions in computational biology:

• Developed methods for uncertainty-guided optimization in VAE latent spaces and efficient training of large ML models
• Created deep generative models for predicting of the effects of genetic mutations in humans with accuracy matching experimental methods, and for identifying viral mutations likely to escape immunity based on prepandemic information only
• Built hybrid architectures combining protein language models with generative models of evolutionary data, achieving state-of-the-art performance in fitness prediction (detailed in this GSK.ai blog post)
• Pioneered the first joint generative models of sequence and experimental measurements, with applications in protein design
• Conceptualized and led the development of ProteinGym benchmarks to robustly evaluate the capabilities of protein fitness and design models

A while back

In what almost feels like a different life, I served as a Senior Engagement Manager at McKinsey & Company in New York, advising healthcare and pharmaceutical clients on Digital & Analytics strategy. During my six-year tenure there, I learned how to lead teams in fast-paced projects, identify and execute on high-leverage opportunities, communicate effectively across interdisciplinary teams, and solve complex, ambiguous problems. While I valued the experience gained across 30+ client engagements spanning 10+ countries, my passion for scientific discovery ultimately drew me back to academia.

Prior to McKinsey, I gained solid foundations in Applied Mathematics and Physics through undergrad studies at École Polytechnique, and a MS at Columbia University.