About the Role
Build the reinforcement-learning environments and harnesses that let frontier models practice real financial workflows, and be graded on them.
A full-time role based at Qofi's New York office, working alongside operators from top investment banks and private equity firms and engineers from leading AI labs.
Qofi builds the highest-precision financial-reasoning data in the world, and the platform and deployments that put it to work inside institutions.
High ownership from day one, in a flat structure where ideas move from whiteboard to production without committees.
Base salary plus meaningful equity and full benefits. Range is a placeholder pending confirmation.
On-site at Qofi's New York office, alongside the team building the data engine.
Full-time. Rolling start, the team moves quickly once there's a fit.
What You'll Do
Design and implement RL environments that model real finance workflows, from spreadsheet construction to multi-step deal analysis.
Build verifiers, rubrics, and grading harnesses that score model behavior against institutional standards.
Develop the tooling and infrastructure to run, evaluate, and iterate on environments at scale.
Work directly with AI lab partners and internal experts to define new tasks and reward signals.
What We're Looking For
3+ years of software engineering experience, with strong Python and systems fundamentals.
Experience with reinforcement learning, agent evaluation, or simulation environments, or strong aptitude to learn fast.
Comfort building and operating distributed systems and data pipelines.
Interest in how frontier models learn to reason about complex, real-world tasks.
Hiring Process
Send a resume and a short note on why this role, about five minutes.
A conversation about your background, the role, and what you're looking for.
Technical and team-fit conversations with the people you'd work with directly.
A clear decision quickly, with an offer and a start date that works.