Sapien AI · Remote · US or Canada
CAD $200,000 to $250,000 base salary · 0.75% to 1.5% equity
Sapien is becoming a research and model development company, starting with AI judges for robotics, world models, and physical AI. We post train open weight models to match the decisions and reasoning of our customers' experts, extending their judgment across far more data while they keep setting the standards and checking the results.
As our first research hire, you'll build a public benchmark, develop better judges, and research how their feedback can improve existing robotics and world models. Our ambition is to take what we learn and build models of our own.
You'll work directly with our founders and engineering team, run experiments from week one,
and build the research team as the work grows.
What You Bring
- You're currently enrolled in a PhD program or have completed a PhD in machine
learning or a related field.
- You've owned and published original research in ML evaluation, reward modeling,
post training, or a closely related area.
- You've built a benchmark others use or an evaluation suite that informed what
shipped, with methods tested across multiple datasets.
- You've trained a model that reached production and can run experiments from
data preparation through training and evaluation.
- You can set research priorities, explain your decisions, and take a research
question through to a defensible result. Robotics experience is welcome and optional. Experience leading researchers is helpful, and doing the research yourself is central to the role.
What Success Looks Like
- First week: Work with our robotics data, experts, and current judges, and run your first
experiments.
- First 30 days: Have a benchmark design, baseline comparisons, and initial results.
- First 90 days: Have a working benchmark and trained judges for priority customer tasks, with
held out results against robust baselines. Run an initial experiment measuring how judge feedback affects downstream model performance.
How to Apply
Email
[email protected] with your résumé or research profile and one short example of a benchmark, evaluation suite, or model you built. Tell us what you owned, one decision that mattered, and what evidence showed it worked.
📌 Research Lead (Toronto)
🏢 Sapien
📍 Toronto