Divya Nori
@divnori
Followers
582
Following
356
Media
10
Statuses
56
CS PhD @Stanford @ArcInstitute | @MIT alum | AI x Bio
Palo Alto, CA
Joined July 2020
Happy 🦃 Thanksgiving weekend! 🍂 This year, we cooked up a new recipe for juicy fact-storing MLPs. Instead of picking apart trained models, we asked: Can we construct fact-storing MLPs from scratch? 🤔 Spoiler: we can & we figured out how to slot these hand-crafted MLPs into
8
46
335
Announcing onepot — making small molecules faster and cheaper Read more at https://t.co/bzgYekesAm
0
4
6
What if we could autocomplete DNA based on function? Today in @Nature, we share semantic design—a strategy for function-guided design with genomic language models that leverages genomic context to create de novo genes with desired functions.🧵 https://t.co/P5qVJB3qIY
16
154
590
Today in @Nature, in work led by @aditimerch, we report the ability to prompt Evo to generate functional de novo genes. You shall know a gene by the company it keeps! 1/n
7
103
543
We are actively recruiting for two positions at the interface between biology and generative design. Backgrounds of particular interest are in protein biochemistry/evolution and synthetic genomics/biology. Please consider joining us! 1/n
8
92
348
Congrats @divnori and the @ValthosTech team on the launch🚀! Couldn't imagine a more competent team to tackle this increasingly important problem for American biodefense 🇺🇸.
Valthos builds next-generation biodefense. Of all AI applications, biotechnology has the highest upside and most catastrophic downside. Heroes at the frontlines of biodefense are working every day to protect the world against the worst case. But the pace of biotech is against
1
1
11
One of the greatest opportunities in AI safety and security is the chance to help support the creation of new industry verticals that help maximize the benefits and minimize the risks. The technology is moving fast. One of the best ways to keep up is with more technology, more
Valthos builds next-generation biodefense. Of all AI applications, biotechnology has the highest upside and most catastrophic downside. Heroes at the frontlines of biodefense are working every day to protect the world against the worst case. But the pace of biotech is against
137
141
692
for the past few years, pushing the frontier of generative biology has been my driving goal. this summer as part of the team @valthostech, i realized that as these capabilities advance, it’s more important than ever to deploy and operationalize them to safeguard humanity against
Valthos builds next-generation biodefense. Of all AI applications, biotechnology has the highest upside and most catastrophic downside. Heroes at the frontlines of biodefense are working every day to protect the world against the worst case. But the pace of biotech is against
13
7
86
This Tues (10/14) at 4PM ET, @samuelhking will present "Generative design of novel bacteriophages with genome language models" Paper: https://t.co/ylXGC3aZtV Then, on Wed (10/15) at 4PM ET, @r_krishna3 will give an early-career talk on his research!
biorxiv.org
Many important biological functions arise not from single genes, but from complex interactions encoded by entire genomes. Genome language models have emerged as a promising strategy for designing...
1
8
46
The ability to design antibodies against any protein of interest has major implications for medicine, biotech, and basic science. Today, we introduce Germinal, a pipeline for epitope-targeted de novo antibody design achieving 4–22% success rates with efficient experimental
23
157
769
Many of the most complex and useful functions in biology emerge at the scale of whole genomes. Today, we share our preprint “Generative design of novel bacteriophages with genome language models”, where we validate the first, functional AI-generated genomes 🧵
40
328
1K
6 months, 25 million revenue agents & 3 trillion tokens later... Rox is now globally available 🌎 Just as coding agents 10x’d engineering, revenue agents 10x customer work. With Rox, humans are evolving to orchestrators while agents manage the end-to-end customer lifecycle.
94
88
648
🧵(1/n) Our study of the maternal gut microbiome and preterm birth is finally out in @cellhostmicrobe. It was a great honor to be part of the team and to be mentored by @MiaoZelei and Prof. Zheng (@zheng_jusheng) during my undergraduate studies (as part of my thesis work :)).
1
9
59
Had some downtime before starting my PhD and decided to get back into creative writing. Ended up with a short essay on taste—in writing, in research, and why it feels hard to cultivate yet more important in the AI era. Sharing if you’d like to read :) https://t.co/UqSQbqdyKe
1
1
8
💻 BECraft Code: https://t.co/4PCoQJze2R Very grateful to work with Anisha Parsan, @CarolineUhler, and @WengongJin. Huge thank you to the original BindCraft creators, @MartinPacesa @sokrypton @befcorreia, for building a powerful foundation and making the code easy to extend
1
0
3
Beyond hallucination, pTMEnergy shines in virtual screening. It outperforms supervised, physics-based, and structure-based baselines at classifying miniprotein binders and RNA aptamers.
1
0
2
By integrating pTMEnergy into the hallucination loop, we get BECraft, a drop-in extension to BindCraft that backprops through AlphaFold2 using this new energy. BECraft outperforms BindCraft, RFDiffusion, and ESM3 across 8 diverse targets 🚀
1
0
1
We fix this with pTMEnergy 💡 It reinterprets AlphaFold’s pAE logits as an energy-based model (EBM) which: ✅ Reflects the true statistical likelihood of the binder-target complex ✅ Enables dense gradients across the entire interface
1
0
1
BindCraft has become a go-to tool for protein binder design, using AlphaFold2 hallucination + ipTM-based backpropagation to optimize sequences. But ipTM has limitations: ⚠️ It's a heuristic, not a proper likelihood ⚠️ It yields sparse gradients, making optimization tricky
1
0
2