Frances Ding

Frances Ding

Machine Learning Scientist

Prescient Design, Genentech

Biography

I am a Machine Learning Scientist on the Foundation Model (Large Language Modeling) team at Prescient Design, Genentech. My research focus is on foundation models for advancing science. I did my PhD in machine learning at UC Berkeley advised by Moritz Hardt and Jacob Steinhardt, where I worked on algorithmic fairness, interpretability of deep networks, and training set compositions for foundation models in biology.

During my PhD, in the summer and fall of 2022 I was AI resident at Google X working on biological sequence design. Before my PhD, I was lucky to be advised by Sebastian Tschiatschek while at University of Cambridge and Cynthia Dwork and Jeffrey Macklis at Harvard University. I am grateful for support from the Gates Cambridge Scholarship during my MPhil and for support from the NSF GRFP and the Open Philanthropy AI Fellows Program during my PhD.

Education
  • PhD in Computer Science, 2024

    University of California Berkeley

  • MPhil in Machine Learning, 2018

    University of Cambridge

  • BA in Biology, 2017

    Harvard University

Publications

(2025). Generalists vs. Specialists: Evaluating LLMs on Highly-Constrained Biophysical Sequence Optimization Tasks. In ICML, 2025.

PDF Cite

(2023). Cbln1 Directs Axon Targeting by Corticospinal Neurons Specifically toward Thoraco-Lumbar Spinal Cord. In Journal of Neuroscience, 2023.

PDF Cite

(2022). Anticipating Performativity by Predicting from Predictions. In NeurIPS, 2022.

PDF Cite

(2021). Grounding Representation Similarity with Statistical Testing. In NeurIPS, 2021.

PDF Cite

(2021). Retiring Adult: New Datasets for Fair Machine Learning. In NeurIPS, 2021.

PDF Cite

(2020). Representation via Representations: Domain Generalization via Adversarially Learned Invariant Representations. In arXiv, 2020.

PDF Cite