Jeremy Wohlwend Profile
Jeremy Wohlwend

@jeremyWohlwend

Followers
713
Following
86
Media
5
Statuses
50

PhD Student @ MIT CSAIL, former Research Engineer at ASAPP Inc.

Joined June 2014
Don't wanna be here? Send us removal request.
@jeremyWohlwend
Jeremy Wohlwend
5 days
RT @GabriCorso: 📢 Call for proposals: Boltz small-molecule design collaboration! 🧬.Can we help design your ideal molecule? Can you help us….
0
67
0
@jeremyWohlwend
Jeremy Wohlwend
9 days
RT @felix_faltings: Excited to share our new ProxelGen paper!. Completely different from RFDiffusion etc., we generate proteins as densitie….
0
18
0
@jeremyWohlwend
Jeremy Wohlwend
27 days
RT @moritzthuening: @tenstorrent cares about Boltz as well!
Tweet media one
0
2
0
@jeremyWohlwend
Jeremy Wohlwend
27 days
Also cool to see Boltz-2 highlighted on stage today by Rory Kelleher’s keynote at #GTC2025. Read more about NVIDIA’s work from their Blog Post:
0
3
14
@jeremyWohlwend
Jeremy Wohlwend
27 days
All of this translates in end to end speedups that range from 1.5 to over 2x, and even higher for long sequences, with no loss in accuracy!.You can try it out directly on our repo, updating to version 2.1.1.
1
3
12
@jeremyWohlwend
Jeremy Wohlwend
27 days
Some speed and memory measurements I ran on the Pairformer layer. Ran with no auto-tuning on an A6000 card.
Tweet media one
Tweet media two
1
3
10
@jeremyWohlwend
Jeremy Wohlwend
27 days
Two new kernels: triangular attention and triangular multiplicative layer. While there have been efforts to build kernels for the triangular attention, there hasn’t been as much emphasis on accelerating the triangular multiplicative layer. Both support training and inference!.
1
1
5
@jeremyWohlwend
Jeremy Wohlwend
27 days
The team at @NVIDIA has done such amazing work accelerating Boltz-2 through novel CUDA kernels and deploying Boltz-2 as NVIDIA’s NIM! The kernels are live on the Boltz repo, and you can run the model with 2x training & inference speedup and large memory savings!🧵#cuEquivariance
Tweet media one
4
14
106
@jeremyWohlwend
Jeremy Wohlwend
27 days
RT @NVIDIAHealth: Did you know the NVIDIA #cuEquivariance library can now accelerate Triangle Attention and Triangle Multiplication operati….
0
5
0
@jeremyWohlwend
Jeremy Wohlwend
28 days
RT @AIHealthMIT: A remarkable turnout! Thank you to the 400+ attendees who came to listen to the first-ever Boltz-2 presentation in-person….
0
5
0
@jeremyWohlwend
Jeremy Wohlwend
1 month
Boltz-1 was always a stepping stone. It’s great to finally share what we’ve been up to since. Boltz-2 puts binding affinity prediction to work—so we can run virtual screens without breaking a sweat. A few other goodies too, go check it out!.
@GabriCorso
Gabriele Corso
1 month
Excited to unveil Boltz-2, our new model capable not only of predicting structures but also binding affinities! Boltz-2 is the first AI model to approach the performance of FEP simulations while being more than 1000x faster! All open-sourced under MIT license! A thread… 🤗🚀
0
10
75
@jeremyWohlwend
Jeremy Wohlwend
1 month
RT @GabriCorso: Excited to unveil Boltz-2, our new model capable not only of predicting structures but also binding affinities! Boltz-2 is….
0
418
0
@jeremyWohlwend
Jeremy Wohlwend
2 months
Thrilled to release Boltz-1x! The physical quality of poses was a recurring piece of feedback we received. With Boltz-1x, we use inference-time steering to remove clashes, respect chirality and other physical properties. A few other goodies too. Work led by the amazing @NoahBGetz.
@GabriCorso
Gabriele Corso
2 months
🚀 Excited to release a major update to the Boltz-1 model: Boltz-1x!. Boltz-1x introduces inference-time steering for much higher physical quality, CUDA kernels for faster, more memory-efficient inference and training, and more! 🔥🧵
Tweet media one
0
21
112
@jeremyWohlwend
Jeremy Wohlwend
3 months
RT @TheMarkFdn: 🤯Mind-blowing keynote by @BarzilayRegina on the application of AI in cancer research! 🤖The future is here. @MIT @AIHealthMI….
0
4
0
@jeremyWohlwend
Jeremy Wohlwend
5 months
RT @peholderrieth: New paper out!. We introduce “LEAPS”, a neural sampling algorithm for discrete distributions via continuous-time Markov….
0
29
0
@jeremyWohlwend
Jeremy Wohlwend
5 months
RT @Itamarchinn: Excited to see our work published in @ScienceMagazine on generating novel disordered proteins with PLMs that selectively l….
0
34
0
@jeremyWohlwend
Jeremy Wohlwend
5 months
RT @AIHealthMIT: As lung cancer rates rise in nonsmokers, a troubling trend for the #1 deadliest cancer, we hope more hospitals will adopt….
0
8
0
@jeremyWohlwend
Jeremy Wohlwend
7 months
RT @GabriCorso: The recording of our seminar on Boltz-1 with @mmbronstein and @vant_ai is now public on YouTube!
0
26
0
@jeremyWohlwend
Jeremy Wohlwend
7 months
Check out our repo for all code, data, and instructions: our updated manuscript: and join the Boltz Slack community: Merry Christmas from the Boltz team with @GabriCorso and @pas_saro!.
0
9
34
@jeremyWohlwend
Jeremy Wohlwend
7 months
Finally, to encourage transparent benchmarking in the community, we released a full set of evaluations on PDB test set and CASP15 for Boltz-1, Chai-1 and AlphaFold3! All the instructions at
Tweet media one
1
4
13