F1servicecentre

About the company
Our client builds the training data and evaluation infrastructure that frontier AI labs use to make their models better. They work with the world’s leading labs to design high-signal datasets and run rigorous evaluations that go beyond static benchmarks. It is a small, early team (post Series A) where every engineer directly shapes how the next generation of models learns.
About the role
We are looking for a Software Engineer, RL Environments who designs the environments, datasets and evaluation rubrics that frontier models are trained on. You will work hands-on with research teams at top AI labs: experimenting with data collection strategies, diagnosing model failure modes and building the metrics that show whether a model is actually improving. You go from hypothesis to live experiment fast, and your output feeds directly into training runs at scale.
What you will do
<ul>
<li>Design RL environments, tasks and data slices that expose model failure modes in finance, code and enterprise workflows</li>
<li>Build and refine evaluation rubrics, harnesses and reward signals for RLHF and RLVR training pipelines</li>
<li>Model annotator behavior and run experiments to improve specific model capabilities</li>
<li>Develop quantitative frameworks for measuring dataset quality, diversity and impact on model alignment and capability</li>
<li>Create and manage real-world and synthetic data pipelines</li>
<li>Partner with lab research teams to turn their training objectives into concrete data and evaluation specifications</li>
</ul>
What we are looking for
<ul>
<li>Hands-on experience building reinforcement learning environments yourself, not only training agents in existing ones</li>
<li>1-8 years of experience as a software engineer, Python and/or TypeScript</li>
<li>Signals of excellence in software engineering: strong CS degree from a top university in the U.S., Canada or Europe, top-tier companies or research</li>
<li>Currently at an RL / AI-native company, or ML experience at a top VC-backed startup</li>
<li>Bias for action, willing to do difficult and tedious work to get the data right</li>
<li>Nice to have: quantitative frameworks for dataset quality or diversity, founder or early-stage startup experience, side projects or published papers in AI</li>
<li>Able to work on-site in San Francisco</li>
</ul>
What we offer
<ul>
<li>$180,000-220,000 base + equity</li>
<li>Direct impact on how frontier models are trained, working with the world’s leading AI labs</li>
<li>Small early team with real ownership, post Series A</li>
</ul>
👉 Apply here on LinkedIn :)

Software Engineer, RL Environments

Software Engineer, RL Environments

This job is no longer available

Talents

Recruiters

About