Stephen Ra
@stephenrra
Senior Director & Founding Team @PrescientDesign • @genentech
Have you ever wondered how your specialized biomolecule engineering model compares with a general purpose LLM on bio-optimization tasks? Come check out our work at ICML, in the East poster session (#E-2804) happening right now!. #ICML2025
come by to chat about training LLMs to be black-box token sequence optimizers
Have you ever wondered how your specialized biomolecule engineering model compares with a general purpose LLM on bio-optimization tasks? Come check out our work at ICML, in the East poster session (#E-2804) happening right now!. #ICML2025
arxiv.org/abs/2503.20767 we want to select successful algos, formalized as giving a distribution of design labels that achieves a population-level criterion, e.g. >10% of designs beat wild type. AI can predict which algos succeed, but who knows when these predictions are correct.
✨ new work with @jiwoncpark #icml2025! a zoo of algos exists for designing new proteins & molecules w/ AI🧬🧪how do you pick which one to use? our method selects design algos that will achieve user-specified, population-level success criteria w/ high-prob guarantees.👇
i was asked by a few (inc. @YuanqingWang ) what i meant by this earlier tweet, and since i'm pretty busy, i decided to write a blog post to answer the question 🤦
ai for drug discovery looks a lot like machine translation research/development during the cold war.
it's been more than a decade since KD was proposed, and i've been using it all along .. but why does it work? too many speculations but no simple explanation. @_sungmin_cha and i decided to see if we can come up with the simplest working description of KD in this work. we ended…
We're hiring! 🎉 Join our @PrescientDesign @genentech team in NYC! We're hiring a Principal Software Engineer to lead the development of our flagship AI/ML Lab-in-the-Loop platform for therapeutic molecular design 🧬 Drive impact w/ full-stack skills (esp. front-end) as part of…
Come say hi at my #AISTATS2025 poster tomorrow! 📍 Poster Session I #83 🕒 Sat, May 3 at 3pm Semiparametric Conformal Prediction uses copulas and the one-step estimator to build efficient conformal sets over multiple response variables. Joint work w/ Rob Tibshirani, @kchonyc
📢 Excited to present our poster at the #ICLR2025 @gembioworkshop! I'll be introducing TherAbDesign - a novel, sequence-based framework that efficiently optimizes antibodies toward therapeutically relevant biophysical properties. [1/4]
I will be presenting this work tomorrow at #ICLR2025 at 10 am stop by to know how to build protein language models and use them to design proteins with new properties!
[1/n] Does AlphaFold3 "know" biophysics and the physics of protein folding? Are protein language models (pLMs) learning coevolutionary patterns? You can try to guess the answer to these questions using mechanistic interpretability. But the thing is, more often than not, we know…
We are presenting 5 papers @iclr_conf on AI4Bio / BioML: Interpretable LLMs, all-atom latent diffusion, Agents, & guided sampling from foundation models 🧵1/
ascpt.onlinelibrary.wiley.com/doi/10.1111/ct… i co-authored a short tutorial-ish paper about LLMs on drug discovery. too many things have happened after we wrote it in the last 6 months though, lol
🚨New post on the @berkeley_ai blog! Hopefully more accessible for a general ML audience: - Protein structure prediction vs. generation - All-atom protein generation as a multimodal generation problem - Our work on PLAID! With @nc_frey @PrescientDesign ⬇
Great work and collaboration led by Hanchen (@hcwww_) on agents for spatial genomics/biology! 👇
🌟Meet SpatialAgent: an AI agent for spatial biology (a key tech to understand cancer and develop therapies) It rocks from experiment design to data analysis to hypothesis generation. It supports multi-organs, spans species, and can work alone or team up with human scientists🧑🔬
We have an opening in our group within @PrescientDesign, @genentech at the intersection of machine learning, molecular dynamics, and structure-based drug design. roche.wd3.myworkdayjobs.com/ROG-A2O-GENE/j…
Reliable algorithm selection for machine learning-guided design @PrescientDesign @Genentech 1. This paper introduces a novel method for selecting machine learning-guided design algorithms that can reliably generate desired outputs meeting user-defined success criteria, such as…
new blog post with @amyxlu ! we tell the story of how we figured out latent diffusion for all-atom protein co-generation, with our methods CHEAP and PLAID. link 👇