There are no interpretability workshops ICLR 2026, so stop by Aya Abdelsalam Ismail 's poster (#504) during Poster Session 5 and we'll show you how to build interpretable Language Models by design.

a lot of people think binder design is the bottleneck for drug discovery, but developability is arguably where wetlabs struggle more due to the cost of the assays. swing by Amy's poster to learn more!

long on lobster

Sequencing should be a human right for those with possible rare disease. Clinical care should be -> start with maximal data and work backwards vs. (the current) start with minimal data -> then slowly add more, letting the disease progress

We are hiring a PhD research intern at FAIR w/ Mark Ibrahim Kamalika Chaudhuri to start this summer or Fall! Potential topics: trustworthy and reliable LLMs, multi-modal LLMs and agents, post-training, reasoning, with a focus on open science/sharing our findings in the paper at the end

We're hiring! Come work with me, Tim Sterne-Weiler, and the fantastic teams Genentech Comp Bio / Disco Onc & Prescient Design as a postdoc. Link to apply below 👇

AI is revolutionizing drug discovery right now at enterprise scale. It's not a hypothetical anymore 🚀

i think you should be less concerned about how other people are (mis)using AI and more concerned about whether you are using AI to be more virtuous, productive, and empathetic.

AI monitoring is key to responsible deployment. Our #ICML2025 paper develops approaches for 3 main goals: 1) *Adapting* to mild data shifts 2) *Quickly Detecting* harmful shifts 3) *Diagnosing* cause of degradation 🧵w/ Xing Han 韩星 Anqi (Angie) Liu Suchi Saria arxiv.org/abs/2505.04608

The two most common questions I get via cold email are 1) what should I work on, and 2) how do I get a job doing research, engineering, etc.? I wrote a new post summarizing my advice for people trying to enter Biology+ML, from undergrads to early career. 👇 (link below)

I recently had the pleasure of speaking with my graduate alma mater NYU Center for Data Science about my personal experience building a lab-in-the-loop system for therapeutic antibody lead optimization with my amazing colleagues Genentech. Active learning is a leap of faith!

i was asked by a few (inc. Yuanqing Wang ) what i meant by this earlier tweet, and since i'm pretty busy, i decided to write a blog post to answer the question 🤦

If you, like me, felt frustrated after reviewing 6 papers for NeurIPS Conference and thought "there must be a better way" then this post is for you. Link 👇

& “WATCH: Adaptive Monitoring for AI Deployments via Weighted-Conformal Martingales” by Drew Prinster @ ICML, Xing Han 韩星, Anqi (Angie) Liu, & Suchi Saria proposes a weighted generalization of conformal test martingales: arxiv.org/abs/2505.04608 (5/5)

There is so much deep engineering work that is far upstream of drug discovery and in vivo systems. If you have a strong technical background you can have a very high impact by contributing to things like sequencing instruments, imaging, lab robotics and so on.

come by to chat about training LLMs to be black-box token sequence optimizers

LLMs vs Protein Design Tools: Who Wins? if you missed them ICML Conference, here's the poster for work led by Angelica Chen & Samuel Stanton that answers this question! Link to paper below 👇

🚨 We’re hiring at Frontier Research Prescient Design Genentech! Work on statistical ML for drug discovery: - Bayesian inference & decision theory - Uncertainty quantification - Multi-objective optimization DMs open, feel free to reach out directly! 🔗 tinyurl.com/4bm5st9v