Incoming PhD @Stanford CS | Prev @UCBerkeley EECS | AI Bio

Joined May 2020
1 Photos and videos
Affinity maturation is how naive antibodies evolve into strong binders, but most antibody LMs ignore it. @stephenzlu and I built CoSiNE to learn this, beating antibody LMs on VEP and reframing design as guiding evolution, not de novo generation. Excited to present at ICML!
Antibody LMs learn what looks antibody-like, but not how selection turns naive germline antibodies into strong binders. @aakarshv1 and I are excited to share CoSiNE, a model that learns this germline-to-mature process for variant effect prediction and antibody design. (1/8)
1
7
44
5,180
Aakarsh Vermani @ ICML reposted
Heading to ICML! ✈️ @aakarshv1 and I will be presenting this work on Thursday during poster session 8. Come say hi at poster #808!
Antibody LMs learn what looks antibody-like, but not how selection turns naive germline antibodies into strong binders. @aakarshv1 and I are excited to share CoSiNE, a model that learns this germline-to-mature process for variant effect prediction and antibody design. (1/8)
3
24
2,757
Aakarsh Vermani @ ICML reposted
We are excited to announce that LatchBio has acquired TwentyTwo, a YC S25 team building AI infrastructure for biosecurity. As AI makes biology easier to engineer, the safeguards around these models stop being an afterthought and become core infrastructure. Security has to improve as fast as capabilities to allow science to progress. Within our lifetimes, AI will put the ability to engineer a pandemic within reach of a single bad actor. What once took a nation-state and years of work is collapsing toward an undergraduate and a few weeks in the lab. That world is arriving faster than most AI labs are prepared for. However, in the right hands, these models will accelerate science at a pace never seen before. Scientists will use these models to cure diseases that families have fought for generations and catch outbreaks before they spread. We cannot wall them off from the tools that augment their work. The problem is that the line between the two is hard to draw. Studying SARS-CoV-2 or Ebola to prepare for the next pandemic is very similar to what a bad actor would do to cause one. TwentyTwo builds intelligence defense systems that distinguish this line and apply safeguards or grant capabilities accordingly. Harmon Bhasin, Evan Seeyave, and John Wang are exceptional researchers who care deeply about building the future of life sciences responsibly. The three of them will join Latch as Members of Technical Staff and lead Latch Biosecurity. I am grateful to be working with them. Expect many more announcements from them over the coming days. We are hiring biosecurity engineers. If this is the kind of work you want to do, reach out.
13
14
98
19,353
Aakarsh Vermani @ ICML reposted
A sparse autoencoder aims to learn a dictionary of interpretable features from a model's activations — but a lot can come out "dead," never firing once. On some models this is rare; on others, >70% die even with fixes like AuxK. We went down a rabbit hole to understand why...
5
36
159
39,760
Aakarsh Vermani @ ICML reposted
Monday Starkly Speaking: @YeqingLin_ will present "Fast and Ultra-Capable Protein Design with Genie 3" biorxiv.org/content/10.64898… On Zoom 12pm ET / 6pm CEST: hannes-stark.com/starkly-spe…
1
6
47
8,198
Aakarsh Vermani @ ICML reposted
🍹Weekend project: I (with help from claude - budget $40), was able to get OpenFold3 weights running inside AlphaFold3 codebase (jax). Works for proteins, ligands, rna/dna (1/3)!
AF3 source code is now open source (aka Apache 2.0)! 😎 Though not the weights... 🙃 github.com/google-deepmind/a…
9
62
397
26,892
Aakarsh Vermani @ ICML reposted
So much of protein design comes down to filtering thousands of samples. Excited to share Promera, especially the binder discrimination results and our interface contact score (iCS), where we explore scoring beyond structural confidence metrics! Also check out our in silico design case studies, including VHHs predicted to stabilize the active like I121/F282 toggle residue state of β2AR! Code: github.com/bjing2016/promera
@mihirbafna14 and I are excited to introduce Promera, a co-folding and design model with • best-in-class binder filtering • nanobody design with in-silico success rates matching hallucination • case studies on hantavirus epitope targeting and GPCR agonism (1/8)
17
120
26,153
Aakarsh Vermani @ ICML reposted
I should be able to build a bioweapon on the $200 a month plan. For $20, yeah, it seems reasonable to ban it
25
108
3,715
109,369
Aakarsh Vermani @ ICML reposted
academics are unprepared for the coming world where much scientific progress is majorly a function of inference compute. whether OpenAI points the Eye of Stargate at your particular field will decide its acceleration. talent will leach away into the labs. it's already begun
10
45
673
42,803
Had a lot of fun working on this project! Particularly excited by our "convergent hotspots" heuristic for binder design, which iteratively refines the target interface at inference time. Simple yet effective! Try it out: github.com/aqlaboratory/geni… Preprint: biorxiv.org/content/10.64898…
Equivariance is dead😢 Or is it?😈 Genie 3 is out! Our latest protein design model yields SoTA results for binder design & motif scaffolding, greatly improving on BindCraft & Proteina-Complexa It does so using all-atom SE(3)-equivariance on a branched polymer representation👇
1
8
520
Note that the talk will also be webcast over Zoom, and the link will be sent out on the day of the talk to those who RSVP!
1
152
We'll be closing out this semester's Berkeley BioML Seminar on 4/28 with a talk from @antoinekoehl on PEINT, a powerful deep learning framework for both phylogenetic inference and protein engineering. Sign up below! luma.com/3fbbftwy
2
2
6
2,677
Aakarsh Vermani @ ICML reposted
We're excited to announce the fourth Berkeley BioML seminar of the semester happening next Tuesday 4/7! Join us for a talk by Shreshth (@shreshth_gandhi) from @tahoe_ai on Scaling Perturbation-Trained Single-Cell Foundation Models! luma.com/njhd2xga
5
9
1,550
Aakarsh Vermani @ ICML reposted
Wrote a deep dive on implementing a language model from scratch in JAX and scaling it with distributed training! If you’re coming from PyTorch and want to see how the same ideas look in JAX, or just want a hands-on intro to distributed training, check out this blog post: chuyishang.com/blog/2026/jax… Comes with code an assignment and test cases so you can follow along!
8
65
602
33,872
Aakarsh Vermani @ ICML reposted
We're excited to announce the third Berkeley BioML seminar of the semester happening next Tuesday 3/2! Join us for a talk by Vivek Natarajan (@vivnat ) from GDM on advancing science and medicine with collaborative AI agents. luma.com/532xzjiu
6
11
1,783
Aakarsh Vermani @ ICML reposted
Can we simulate realistic evolutionary trajectories and “replay the tape of life”? In this work, we propose a flexible, generalizable framework for modeling how the entire protein seq evolves over time while capturing complex interactions across sites. 1/n doi.org/10.64898/2026.02.19.…
5
82
288
34,197
Aakarsh Vermani @ ICML reposted
We're excited to announce the second Berkeley BioML seminar of the semester happening next Tuesday 2/17! Join us for a talk by Kenny Workman (@kenbwork) from LatchBio about the performance of agents for spatial biology analysis. luma.com/f3xa3dst
2
4
3,583
Another semester, another Berkeley BioML Seminar! We're excited to kick off this semester with a talk from @jeffruffolo from Profluent Bio on designing proteins with language models. Join us Monday night! luma.com/dknwfirh
2
11
862