Apply with the biggest problem in your field. If you’re selected, you spend ten weeks pushing it forward with frontier AI, then show the results at Demo Day.
Pick one result to move this week. A concrete build, not a vague intention.
Researchers and operators from AI labs and science-AI companies share the workflows they actually use and ship.
Work from a skills library that’s yours to keep, with perks from our AI partners and 1:1s with us to unblock broken agent loops.
YC-style group office hours: show what changed, get pushback from other fellows, and take ideas from one lab to another.
Bring the problem and your knowledge of the field. We’ll work alongside you for ten weeks to accelerate your research.
Learn from the frontier with us. We help you turn new possibilities into a concrete research goal.
We help you choose a hypothesis or use case where AI can make a real difference, then set a concrete goal.
Guest speakers from frontier AI labs and startups show what’s possible. Explore real use cases and find new ideas for your research.
Meet previous guest speakersWe help you build and get unstuck. Demo at office hours each week, and we’ll help you choose the next step.
AI credits to test more ideas, run larger analyses, and keep experimenting.
You don’t start from scratch. Adapt reusable AI skills and workflows to your research, and keep them after the fellowship.
Get 2 hours every week of hands-on help to build AI tools and workflows for your research, solve technical problems, and keep moving.
Work with us to sharpen your goals, think through roadblocks, and stay accountable to your research ambitions.
Present your progress at the end of every week. Get thoughtful critique, see what others are building, and take new ideas back to your work.
Present what you built, discovered, and will pursue next to scientists, AI researchers, and investors. We help you prepare your talk.
See Cohort 1 outcomesYour research, under your name. We help you write up your problem, approach, and findings so others can read your work and build on it.
See Cohort 1 Case StudiesYour 2027 journey, from the first conversation to your work in public.
Tell us about your scientific problem and ambition.
Explore your goals with us. Fellows are selected in February.
Meet your cohort and set your first research goal, then continue online.
Build on your own research with talks, hands-on support, and weekly demos.
Return to share what you built and discovered with a wider audience.
Turn your work into a published case study with our help.
* Exact dates for interviews and the program will be announced.
Guest speaker
Reactorfield
Professor of Immunology at The Jackson Laboratory
Guest speaker
Reactorfield
Director, Healthcare & Life Science AI Startups at NVIDIA
Guest speaker
Reactorfield
Founder at Protocol Labs
Guest speaker
Reactorfield
Co-founder at Yukon, Founder of EigenLayer
Guest speaker
Reactorfield
Co-founder & CEO at Rowan Scientific
Guest speaker
Reactorfield
Co-founder & CEO at Hyperspell
Guest speaker
Reactorfield
Member of Technical Staff at LatchBio
Guest speaker
Reactorfield
Co-founder & CEO at AnswerThis
Guest speaker
Reactorfield
Founder & CEO at OpenWork
Guest speaker
Reactorfield
Co-founder at muni
Guest speaker
Reactorfield
Founder & CEO at Impulse AI
Guest speaker
Reactorfield
Co-founder at muni
Guest speaker
Reactorfield
Yukon
Guest speaker
Reactorfield
Founder at Tuva AI
You’re pursuing a PhD, deep in a postdoc, or leading a lab. Your days revolve around experiments, papers, and a scientific question you can’t leave alone.
You do R&D in biotech, materials, energy, or another scientific field. You know the satisfaction of a good experiment, and the questions that come with an unexpected result.
You’re building a company around hard science. You move between research questions, early prototypes, and the challenge of turning a discovery into something people can use.
March to May 2027. Applications are open now and reviewed on a rolling basis until the January 24, 2027 deadline. Interviews run in February, the program starts in March, and it ends with Demo Day in May.
Scientists, engineers, and researchers across domains: academic, industry, or deep tech startups. Teams are encouraged to apply together: the program is for the people doing the work, not only the person with the title.
Apply with the biggest problem in your field. We review every application and interview the strongest, usually one or two calls before a fast decision. We select for the size of your scientific ambition and your urgency to put AI to work on it.
Ten weeks, and every week counts. Each one has a goal, a session with researchers from AI labs, 1:1 time with us and YC-style group office hours, plus serious build time on your own research. Fellows who protect the hours get the results; there is no passive version of this.
Hybrid. Live sessions run online throughout the program, with in-person time for the cohort.
No. Reactorfield is free and non-dilutive, supported as a public-good initiative to accelerate science.
No. You keep all your IP. We don't train on your data, it's all yours.
No. You need depth in your own science and a willingness to rebuild how you work. We close the gap between what frontier models can do and how you use them. That is the whole point of the program.
Ten weeks of results on your own problem, shown at Demo Day and written up in a published case study. A skills library and workflows that run on your questions and your data, the judgment to keep them current as the tools change, and a cohort of peers across deep tech who share what works.
Reactorfield is like magnetic field and gravitational field. A field that accelerates and transforms what interacts with it. We engineered Reactorfield as a space that catalyzes reactions and triggers inflections. We tip our hat to physics; the laws of the universe are our only true bounds. (and also because we're nerds)
Applications are reviewed on a rolling basis until January 24, 2027.
Start your application
