Profluent
@ProfluentBio
Followers
4K
Following
21
Media
75
Statuses
155
Grounded in nature, authored by AI
Emeryville, CA
Joined September 2022
Today we’re announcing $106M in new funding led by Altimeter Capital and Bezos Expeditions. This brings our total to $150M to scale our frontier AI models which make biology programmable. Our frontier models have generated functional proteins (Nature Biotech, 2023), created the
16
37
321
The work will combine Ensoma’s powerful in vivo HSC engineering platform with Profluent’s frontier AI models for protein design to create the next generation of custom-designed editors and intelligent delivery platforms to unlock treatments previously out of reach.
0
0
1
Today we’re excited to announce a new strategic collaboration with @EnsomaBio to advance AI-designed base editors for hematopoietic stem cell (HSC) therapies. The goal: durable, one-time treatments that target the root causes of hematologic and immune diseases.
1
3
14
📅 Thu 4 Dec 4:30 - 7:30 PM PST 📍Exhibit Hall C,D,E #1702 🔗 https://t.co/AVD0TYKMgh 👀 Look for Aadyot, lead ML scientist at Profluent!
0
0
0
The upshot is that larger models generate more viable proteins across more protein families, especially those under-represented in the training data. This unlocks design for therapeutic proteins, like ultra-compact gene editors, that were previously out of reach.
1
0
0
For ProGen3, we curated the world's largest Protein Atlas. We developed compute-optimal scaling up to 46B parameters and trained on 1.5T amino acid tokens. Then we designed a wet lab experiment to evaluate how model scale influences alignment with real data.
1
0
0
#NeurIPS2025 friends! Meet the Profluent team *today* to talk about ProGen3, our generative protein language model that’s solving problems in medicine and agriculture. We're sharing how we built ProGen3 and how we validated it with real wet-lab data.
1
0
4
Or catch us at our spotlight poster on Thursday afternoon to hear the latest on what we're building at Profluent. 📅 Thu 4 Dec 4:30 - 7:30 PM PST 📍Exhibit Hall C,D,E #1702 🔗
0
0
0
Drop us a line on this contact form if you'd like to connect with someone from our team in San Diego this week https://t.co/Cta1qBb5iZ
docs.google.com
Let us know if you'd like to meet with someone from the Profluent team! We're generally interested in chatting about language models for protein design. At NeurIPS, we'll be sharing more about the...
1
0
0
Good morning #NeurIPS2025! Profluent is here and we're hiring! Multiple roles across ML science, engineering, SWE, and protein design. https://t.co/qCStt1IxA3
job-boards.greenhouse.io
Profluent is an AI-first protein design company. Founded in 2022, we develop deep generative models to design and validate novel, functional proteins to revolutionize biomedicine. Based in Emeryvil...
2
0
3
We have multiple team members on site. If you're interested in chatting, contact us here! https://t.co/mdeK20batr
docs.google.com
Let us know if you'd like to meet with someone from the Profluent team! We're generally interested in chatting about language models for protein design. At NeurIPS, we'll be sharing more about the...
0
0
1
Excited for #NeurIPS2025! Join us to discuss language models for protein design ranging from ProGen3 (NeurIPS spotlight), E1 (our open-source encoder model), and more!
1
1
3
First, yay Profluent - I really admire what they do and hope they succeed. Second - folks, that's a C2H2 ZFP there. Pavletich and Pabo Science 1991. ❤️ Pardon me while I go all warm and fuzzy here.
Today we’re announcing $106M in new funding led by Altimeter Capital and Bezos Expeditions. This brings our total to $150M to scale our frontier AI models which make biology programmable. Our frontier models have generated functional proteins (Nature Biotech, 2023), created the
2
3
61
Love this take from @nathanbenaich and @airstreetpress. What works in natural language can work in protein engineering too. With E1, reading a sequence alongside its evolutionary neighbors leads to stronger results. E1 outperforms ESM-2, ESM C, and PoET on ProteinGym and CAMEO.
press.airstreet.com
What works in natural language modeling also works in protein engineering!
new on @airstreetpress: what works in natural language also works in protein engineering! we dive into @profluentbio latest open-source model called E1, which reads an input sequence plus its evolutionary neighbours to outperform ESM-2, ESM C, and PoET on protein design tasks
0
0
12
There’s always been a tension between addressing rare diseases and making the economics work. At Profluent, we believe that AI can be the equalizer, making it possible to design personalized genomic medicines and democratize the field of gene editing. Learn more about our
1
1
7
AI can bring high quality genetic tools within reach for rare diseases. Today, Profluent is partnering with the Rett Syndrome Research Trust (RSRT) to design personalized genomic medicines for Rett syndrome. Our aim is to engineer compact editors that can fit into a single AAV
3
4
22
Oh, and one more thing, we are open sourcing E1 for free research and commercial use. We built it to tackle hard protein design problems at Profluent; so if it is useful for us, it will be useful for you. Paper: https://t.co/L2I19SuffK Blog: https://t.co/cJMcteflwZ Repo:
profluent.bio
We are announcing Profluent-E1, our first Retrieval-Augmented Protein Encoder Model, available under a permissive license for both research and commercial use.
2
4
28
E1 is practically very useful. Hand it a variant and get a principled estimate of how functional it is likely to be, ideal for prioritizing designs. Or generate an embedding for your downstream task. Check out our Github repo for examples:
github.com
Profluent-E1 family of Protein Encoder Models. Contribute to Profluent-AI/E1 development by creating an account on GitHub.
2
1
18