AI-native materials discovery powered by unpublished experimental data
San Francisco, CA
Joined June 2026
- Tweets11
- Following17
- Followers104
- Likes9
83 Sciences retweeted
Our first pre-print is out!
chemrxiv.org/doi/full/10.264…
We spent a big chunk of the summer demonstrating the value of data sitting in filing cabinets and untouched file servers in academia. @83sciences uses this data to drive novel materials discovery.
Huge thanks to Taylor Sparks for the collaboration!
Feel free to reach out to me, @ericriesel, or @YKScience if you're interested in becoming an academic or industry collaborator!
Elon Musk claims we've "exhausted the cumulative sum of human knowledge in AI training", but almost 90% of scientific experimental data never gets published.
83 Sciences discovers groundbreaking materials by training frontier models on published, unpublished, hidden, AND failed experimental data.
In just two months, we’ve already discovered a novel material and drafted a manuscript for publication.
Industrial R&D teams partner with us to find scalable application-specific materials with performance that frontier models can't predict. Academic labs turn decades of unpublished data into new papers, citations, and commercialization revenue with our novel research tools.
#AI #materialsdiscovery
83 Sciences retweeted
Attending YC’s Startup School this weekend was a great reminder that we're building 83 Sciences at the right moment. Young talent is already pushing the boundaries of materials science by conducting research at top research labs. Almost every conversation surfaced the same pain point we saw firsthand: years of valuable experimental data being unfortunately difficult to search, reuse, and learn from, and frontier industries experiencing the physical bottleneck of innovative materials.
To scale, we're hiring a Founding AI Engineer to join us early. If building AI for science sounds interesting, reach out and connect with me at [email protected]
83 Sciences retweeted
This is cool. 83 Sciences is using AI to mine discarded lab data (like failed experiments) for scientific breakthroughs.
The next industrial era will be built on materials discoveries and insights hiding in the 90% of experimental data that never gets published.
Innovate faster, publish more, do more with your data with 83 Sciences
83sciences.ai
🙋♀️We're hiring our Founding AI Engineer at 83 Sciences!
What you'll do: help design insight and discovery extraction models and ship tools for scientists and work closely with the founders
The next industrial era will be built on materials discoveries and insights hiding in the 90% of experimental data that never gets published.
Innovate faster, publish more, do more with your data with 83 Sciences
83sciences.ai