Hiring nowPost-training and preference optimisation

Reinforcement Learning Research Engineer

Also called RL Research Engineer

A Reinforcement Learning Research Engineer improves how a model behaves by training it against interactive signals. They matter because reward gains only count if the model gets better at the real task.

What they do

They build and run the training loop that turns a pretrained model into a better policy. That means writing reward functions, setting up environments or feedback sources, and choosing and tuning the optimisation method. They track experiments closely and look for runs that go wrong. Reward hacking is the usual culprit. They write evaluations that sit apart from the reward, so improvement can be checked independently. Much of the week goes on reading failed runs, adjusting the reward specification, and deciding which result deserves more compute. They work in PyTorch and an experiment tracker, and share findings with researchers and product engineers.

What a strong hire delivers

In the first months, a strong hire delivers a working RL training pipeline that others can rerun and extend. They add a policy evaluation set that is independent of the reward. They produce a clear account of which gains held up and which were artefacts. They leave the team with a reward specification that is written down, tested and understood, plus a shortlist of the next experiments worth running.

How you can tell they are good

Ask them to walk through a training run that went wrong. A strong candidate explains how they noticed, what the reward was really encouraging, and how they fixed the specification. Ask how they know a gain is real. They should talk about held-out tasks, independent success checks and variance across seeds. Be wary of anyone who cites reward curves alone.

Why it is hard to fill

Few people have run reinforcement learning on large models end to end, and the job title varies widely. Many candidates have a strong fine-tuning or classical RL background but not both. Others have trained policies in games or robotics and not on language models. Titles like research scientist, applied scientist and ML engineer all cover this work. As a recruiter, I look for evidence of shipped pipelines, not paper lists.

Where to find them

Most come from post-training teams at frontier labs and model providers, or from applied research groups at well-funded AI start-ups. Others move across from robotics, game-playing agents or recommender systems with a reinforcement learning focus. Strong adjacent candidates are fine-tuning engineers and evaluation specialists who have taken on preference learning and want to work closer to the optimisation itself. Academic labs working on policy optimisation are another source.

Indicative pay

Not enough published evidence yet to give a fair range for this role, which is common for work this new. We benchmark it for you as part of a search.

Interview questions for a Reinforcement Learning Research Engineer

  1. Tell us about a reward function you wrote that did not produce the behaviour you wanted. What happened?

    A strong answer names a specific failure, explains how the model exploited the reward, and describes the fix and how it was verified. Worry if they only describe tuning hyperparameters or claim rewards rarely go wrong.

  2. How do you check that a reward gain reflects real improvement on the task?

    Look for held-out tasks, independent success measures, human spot checks and multiple seeds. A good candidate separates the evaluation from the training signal. Worry if reward curves are their main evidence.

4 more questions, with what a strong answer covers, are in the Role Kit.

Job description you can post

We are hiring a Reinforcement Learning Research Engineer to improve our models through interactive training signals. You will build the pipelines that train our policies and the evaluations that tell us whether they genuinely got better. We value careful experiments and honest results.

What you will do

  • Build and maintain RL training pipelines for large models
  • Design and test reward specifications and feedback sources
  • Run, track and compare experiments across optimisation methods
  • Investigate unstable runs and reward hacking
  • Build policy evaluations independent of the reward
  • Report findings clearly to researchers and product engineers

What you will bring

  • Demonstrated work training models with reinforcement learning or preference learning
  • Strong PyTorch and deep learning experience
  • Evidence of sound experiment design and use of experiment trackers
  • Experience building evaluations that check real task success
  • Ability to explain a failed run and what it taught
  • Clear written communication of results

Nice to have

  • Experience with large language model post-training
  • Background in robotics or agent environments
  • Published or open-source work on policy optimisation

Free to use and adapt. If it saves you time, a link back to this page is appreciated.

Skills that matter

  • Reinforcement learningAdvancedOptimise policies from interaction and reward
  • Preference learningAdvancedLearn behaviour from comparisons and reward signals
  • Deep learningWorkingBuild and diagnose trainable neural models
  • Model evaluationWorkingMeasure capability and failure modes credibly

Other roles in post-training and preference optimisation

Hiring nowPost-Training Research EngineerA Post-Training Research Engineer adapts pretrained models into ones that behave as the company needs. They turn preference data and reward signals…Hiring nowReward Modelling ResearcherA reward modelling researcher builds models that learn what good output looks like from preference comparisons. They give post-training teams a…

Related roles in other families

EmergingReasoning Research ScientistReasoning, planning and searchA Reasoning Research Scientist tests and builds methods that make a model's multi-step inference reliable. They show the company which approaches…Hiring nowPlanning Algorithms ResearcherReasoning, planning and searchA Planning Algorithms Researcher builds planners that choose sequences of actions under hard constraints. Companies hire them so agents and systems…EmergingInference-Time Compute ResearcherReasoning, planning and searchAn Inference-Time Compute Researcher decides how much thinking a model should spend on each problem. They raise accuracy on hard tasks while keeping…EmergingNeuro-Symbolic AI ResearcherReasoning, planning and searchA Neuro-Symbolic AI Researcher combines learned models with symbolic rules and solvers, so systems reason consistently where rules matter. That…

How we wrote this

Role definitions are our own editorial work, checked against occupational classifications, standards, research and employer pages. Titles vary by company, and this is a description of the work, not a job advert.

← Post-training and preference optimisationAll new job families

Start a search

Hiring a Reinforcement Learning Research Engineer?

Tell us the work. We find the people already doing it, whatever their job title says, and you meet them already interviewed for your role.