Nima Alidoust
@nalidoust
CEO and Co-Founder, @tahoe_ai (previously Vevo) , Princeton PhD *15 زن، زندگی، آزادی
Historic day for builders in bio: We’ve open-sourced @vevo_ai’s #Tahoe100M, largest single-cell atlas ever—by a wide margin—as the inaugural contribution to @arcinstitute’s Virtual Cell Atlas, ready for download today. A leap forward for AI models of cells & drug discovery. 🧵
Excited to join @nalidoust and @tfmiller3 on stage next week at @decodingbio and @NYSE’s AIxBio Summit! AI and novel data are reshaping biotech in big ways - can’t wait to dig into the details. See you in NYC! 🧬🚀
Our CTO @LindaBGoodman will be speaking at the @NYSE Decoding Bio AI x BIO Summit next Tuesday in New York. Hope to see you all there! 🧬🧬
On the floor of NYSE for AI x BIO 2025! Thanks @ameekapadia @pablolubroth for the awesome energy and hosting such a talented crowd. 🧬🚀 @nalidoust
A brief musical note: I’m very excited to share that my brilliant wife, a pianist and composer, has just released her debut album featuring works by Iranian women composers, including one of her own. Here is the link in case you’d like to take a listen: navonarecords.com/catalog/nv6747/
I personally find Formation Bio’s story an inspiration for building a biotech brick by brick, with persistence, and with the express purpose of addressing a key bottleneck (not merely a technically novel solution for a solved problem). We have used it as one of the blueprints…
Hi David, don’t disagree with your core message. However, I do want to correct the record that @formationbio always planned to be a full stack Pharma company, differentiated by faster/more efficient clinical trials + drug development. In fact, this was articulated in our seed…
Mmm… so the guy got confused about what a “work wife” means and that’s what led him to cheat on his wife!?
One thought on the Coldplay Tryst that’s important for companies: nothing irks me more than calling a company a family. It is not a family. It is not the same kind of relationship, love or commitment that you have to your kin. You shouldn’t have a “work wife” or treat your job…
Just got my @PerplexityComet invite. *Really* pleasant surprise: it's not an MVP - it works! The assistant doesn't mess up; it does what you ask & the link to my broader context (emails, drive, calendar, search) functions beautifully. Rare for v1 products in this rat race.
This is an incredible program. Founders of any stripe - take a look.
🚨 applications are open for the fifth batch of @conviction's batch program 🚨 Embed is a Schelling Point for the highest quality early-stage founders in AI: offering $150K cash, $500K+ compute, community and a catalyst to start a startup
We love it when others build dev tools for building on Tahoe-100M. Today, we highlight scDataset in our @tahoe_ai blog. Developed by @davide_dascenzo & @sebacultrera, it makes it easier to train DL models on datasets with the daunting scale of Tahoe-100M (see poster in ICML '25)
💙
Accelerating and scaling *human health* is humanity's most important endeavor 🧬
Thrilled to collaborate with Mathew Garnett @Sanger and Parse Biosciences on the new Cancer Plasticity Atlas – a hugely exciting single cell genomics resource to decode tumor adaptation and resistance. 🔗 helmholtz-munich.de/en/newsroom/ne… #CancerAI #Plasticity #Atlas
Watching @SpaceX launches & landings is one of the most thrilling & captivating experiences for my 3-y old. I asked him today, while watching some of the liftoff videos, whether he wants to be an astronaut. “no I wanna be one of the guys who gets to do the countdown” 😀
Applications opening soon for @CivilizationVC’s 2026 Fellowship — our 6th cohort. Designed to bridge the gap from PhD to startup operator or investor, this program empowers the next generation of biotech leaders. Follow here☝️ to stay updated.
#PhD Survey: when a PhD (#biology, #chemistry) joins or founds a startup, what will be their biggest challenge? 👇🎓🧬
Pleasure to chat with my teammate @vivnat to @Nature news, about our work @GoogleDeepMind on AI co-scientist. "What’s it like to work with an AI team of virtual scientists?" 👇 nature.com/articles/d4158…
Let’s see how the New York primary would go if my X followers were the voters. Please rank order your choices.
Amazing to welcome one of the all time greats of AI and NLP to @aixventureshq Link in reply.
Finally had the chance to read @ElliotHershberg's new piece. Very thoughtful. Don't miss it.
What Are Virtual Cells? centuryofbio.com/p/virtual-cell Two months ago, I started working on an essay to answer this question. The goal was to cover a few of the recent research results. Simple as that! Instead, I went down a rabbit hole exploring ideas around cellular simulation…
With more than 18 entries on the leaderboard already, it's great to see the trade-off playing out between the perturbation score and the DE score! We specifically designed these metrics to capture complementary aspects of perturbation response 🧵
Completely agree. Pre-training on quality data for the first time allows for context generalization - > an important milestone in Virtual Cell Not all data is created equal! 🖥️🧬
The result I'm most excited about from Arc's new State model: The ability to generalize on zero-shot out-of-distribution predictions after pre-training on the TAHOE-100M data set. Whereas PLMs have seemingly benefitted less from scaling data and model size, this is an inkling…
Thanks for the shoutout @ElliotHershberg! This was one of the most exciting parts for us as well. We also found that when using cell embeddings, pre-training on Tahoe-100M improved our zeroshot transfer on genetic or signaling datasets!
The result I'm most excited about from Arc's new State model: The ability to generalize on zero-shot out-of-distribution predictions after pre-training on the TAHOE-100M data set. Whereas PLMs have seemingly benefitted less from scaling data and model size, this is an inkling…
Tahoe-100M FTW 💪
The result I'm most excited about from Arc's new State model: The ability to generalize on zero-shot out-of-distribution predictions after pre-training on the TAHOE-100M data set. Whereas PLMs have seemingly benefitted less from scaling data and model size, this is an inkling…