.png)
AI-Native Physicist Fellowship
Inaugural cohort kicks off January 2027
A 12-month fellowship that funds experienced physicists to run their next major research project with frontier AI systems at the core of the process with the aim of building an open playbook for AI-native scientific work.
Commitment
1 Year
Compensation
100k (average) + compute / other resources
Location
Global (remote)
The gap we're funding
Frontier AI is already changing how physics research gets done, but mostly informally, through individual practices that go undocumented. This fellowship makes that knowledge rigorous and shareable.
We fund physicists to pursue a serious research project of their own, on the condition that they treat their AI workflow as part of the deliverable rather than an implementation detail. Fellows document what worked, what failed, and how results were verified. Over time, this builds a public, tested body of knowledge on what AI-native physics research looks like when done well, developed by the researchers doing the work.
Build the methods that will define AI-native physics
What you will do
Pursue your own research project, with frontier AI systems central to the process. Refine your AI-native workflow, verify its outputs, and document the process in detail. By year's end, you will have produced a scientific result suitable for peer review and a public method write-up on how you got there.
What you bring
A strong publication record and the judgment to recognize when a model is hallucinating, drifting on convention, or fabricating a citation. You already use AI tools in your research and can speak concretely about where they fail. You have a specific 12-month project in mind, not an open-ended exploration, and you're willing to document your methods publicly, including what didn't work.
Who should apply
Working physicists past the PhD, including senior postdocs, research scientists, junior faculty, and unaffiliated independent researchers. Late-stage PhD students may apply if their track record shows the same rigor and independent judgment. This program is not a fit for pure ML-methodology projects, "AI for science" proposals that treat AI as a black box, or corporate research submitted under an academic label.
Program Information
Compensation
$75K–$150K for the 12-month term (~$100K average), adjusted for cost-of-living.
Time commitment
12 months, full-time on your proposed project.
Location
Remote, open to physicists worldwide.
Research output
A peer-reviewed submission (paper, preprint, or benchmark) plus a public method write-up.
Compute and research support
Full access to frontier AI platforms (Claude, ChatGPT, Gemini), API credits, compute resources, and free Theo access with priority support.
How selection works
Apply
Submit a research proposal along with a portfolio link or short working-style demo (e.g. a recorded Claude Code or Codex session on a representative problem).
Shortlist
The program lead and advisory committee review applications and select a shortlist.
Interview
Shortlisted candidates interview with the program lead and at least two advisory committee members.
Award
Award decisions are made by the program lead with advisory committee sign-off.
Application Materials
-
Open to physicists globally
-
Senior postdocs, research scientists, junior faculty, and unaffiliated independent researchers are welcome to apply
-
Cohort starts January 2027
-
2-3 page research proposal
-
Include the budget, compute, GPU, data, or any additional support your project would need
OUR TEAM
This is the space to introduce visitors to the business or brand. Briefly explain who's behind it, what it does and what makes it unique. Share its core values and what this site has to offer.
.jpg)
FirstPrinciples
Riley Carter
CONTENT CRAFTER
FirstPrinciples
Maya Ramirez
ANALYTICS ACE
FirstPrinciples
Ali Torel
SOCIAL SUPERSTAR
Harvard
Noah Pattern
Scientific Advisory Board
Columbia
Olivia Chen
Scientific Advisory Board
London Institute
Ethan Remi