About the role
The Research Residency is a paid year of full-time research for strong engineers and scientists who are new to the field. You join a Frontier AI team, take on a real project with a mentor, and finish the year with a body of work you can point to: a paper, a codebase, a trained model or a result that changed what the team does next.
Frontier AI covers the full stack: pretraining, post-training, reasoning, memory, multimodal models, interpretability, alignment and evaluations. Residents are placed where their background fits and where the team has a concrete question to hand over. Several residents have gone on to permanent research roles, and the programme draws on the shared engineering bench, so projects can cross into other divisions.
Expect close supervision at the start and increasing independence. There is a weekly reading group, a written proposal in the first month and a review at the midpoint and the end. We value people who ask a precise question and are willing to say that the experiment did not work.
What you’ll do
- Carry one research project from proposal through experiments to a written result, with a named mentor and weekly check-ins.
- Reproduce a recent paper from scratch, and report honestly where your numbers differ from the original.
- Write clean, tested research code in PyTorch or JAX that other team members can run and extend.
- Learn the team's infrastructure for training, evaluation and experiment tracking, and contribute fixes to it.
- Present your progress to the wider group at set points and respond to critical questions.
- Take part in the reading group, and contribute reviews of internal and external work.
What you’ll bring
- A degree or equivalent experience in computer science, mathematics, physics, statistics or a related quantitative field.
- Strong programming skills in Python and some prior use of PyTorch, JAX or another deep learning framework.
- Comfort with linear algebra, probability and calculus at the level needed to read a machine learning paper and follow its derivations.
- Evidence of independent technical work, such as a project, thesis, open-source contribution or competition result.
- Clear writing and a habit of documenting what you tried, including what failed.
- Curiosity about how large models work and the persistence to stay with a hard problem for months.
Nice to have
- Prior experience in a research or engineering role outside machine learning, such as physics, biology, finance or compilers.
- A completed online course sequence or published reproduction of a deep learning paper.
- Experience with distributed computing or GPU programming.
- Contributions to an open-source machine learning library.
Work you might take on
- 01
Reproduce a published scaling-law study at small scale and extend it to a data regime the original did not cover.
- 02
Build a probe-based analysis of a mid-sized model's internal representations and test what it can and cannot find.
- 03
Create a new evaluation set for a capability the team cares about and benchmark several checkpoints against it.
- 04
Implement and ablate a recent post-training method, then write up which of its claimed gains survive replication.
Compensation
$190K – $220K
The US base salary range for this role. Every offer also includes equity in Clexerl and, when it spins out, in the venture you help build. We tell you the reasoning behind your number.
Benefits
Health, fully covered
Medical, dental and vision for you and your dependants, premiums paid in full.
Two kinds of equity
Equity in Clexerl, plus a stake in the venture you help build when it spins out.
22 weeks parental leave
Fully paid, for every parent, however your family arrives.
Rest that sticks
Flexible time off with a four-week minimum we actually check on, plus a company-wide week in December.
Learning and compute
A yearly budget for courses, conferences and books, and compute credits for side projects.
401(k) with a 4% match
Matched from your first paycheque and vested immediately.
How we hire
Apply 5 minutes
Send links to what you have built and a few lines on why this role. No cover letter needed.
Intro call 30 minutes
A conversation with the hiring manager about the work, the team and what you want next.
Work sample 2–4 hours
A paid take-home or a pairing session on a real-shaped problem. You pick the format.
Studio day One day
Meet the team, go deep on your work and ours. Remote or in person.
Offer Within 5 days
A written offer with the full range, the equity and the reasoning behind both.
Opens an email with a short template. Links beat résumés.