For engineers
Designs evals and RL pipelines so quality is measured, not guessed. Here is what the role is, what it pays, and what is open now.
An eval engineer makes model quality measurable. They build the harnesses, datasets, and scoring that turn a vague sense of "this feels better" into a number you can move on purpose.
On the RL side, they design the data and reward pipelines that improve a model. Without good evals, RL is guessing with extra steps, so the two jobs sit close together.
In the US, total compensation for strong eval and rl engineers usually lands around $190k to $280k, higher at frontier labs and with equity at startups. We do not post a role we would not take ourselves.
Nothing open under this track this week. If you build this way, send us what you have built and we will keep you close.
They build the evals that measure model quality, so teams can tell whether a change made things better or worse instead of guessing.
Because most AI teams ship faster than they can measure. An eval engineer is what turns shipping into improving.
In the US, total compensation usually lands between $190k and $280k, higher at frontier labs.