Research Scientist, Science of Post-Training and Reinforcement Learning

Snapshot

We are starting a small team aimed at building a real science of post-training for agents. This involves reinforcement learning for LLM-based systems, rigorous experimentation, and a focus on scaling, evaluation, and the practical details that make methods work.

This Research Scientist role is intentionally hands-on. The core loop is: form a hypothesis, implement it, run strong experiments, analyze what happened, and decide what to do next. We care about research that holds up over time, not just incremental wins.

About Us

Artificial Intelligence could be one of humanity’s most useful inventions. At Google DeepMind, we’re a team of scientists, engineers, machine learning experts and more, working together to advance the state of the art in artificial intelligence. We use our technologies for widespread public benefit and scientific discovery, and collaborate with others on critical challenges, ensuring safety and ethics are the highest priority.

The Role

You will work closely with Ian Osband and the team on research around post-training for agents and LLMs, including practical RL methods and evaluation. This is not a theory-only role; you should expect to implement code, run experiments, and own results end-to-end. Success in this role is defined by whether the team learns faster and whether the work produced is crisp, honest, and high-quality.

Key Responsibilities

•
Propose and test research hypotheses in post-training and RL for agents/LLMs.

•
Implement algorithm ideas and run end-to-end experiments, including setup, execution, analysis, and iteration.

•
Design evaluations and ablations that answer real questions and change minds.

•
Analyze results carefully, including debugging and failure analysis.

•
Communicate clearly through plots, writeups, and paper-ready narratives and figures.

•
Collaborate closely with engineering and research partners to keep the team aligned on findings and strategy.

•
Contribute to a culture of first-principles thinking, high standards, and direct, constructive feedback.

About You

In order to set you up for success as a Research Scientist at Google DeepMind, we look for the following skills and experience:

•
A research track record in ML/RL, demonstrated through publications or high-quality projects.

•
Strong implementation ability and comfort working in research codebases.

•
Evidence of owning experiments end-to-end, including analysis and interpretation.

•
Strong communication skills and a bias toward clarity and honesty regarding results.

•
High agency and drive: You push projects forward, prioritize effectively, and take initiative.

•
PhD in ML preferred, or equivalent practical experience.

In addition, the following would be an advantage:

•
Experience with RL for sequence models, post-training, preference-based learning, or agentic systems.

•
Experience with modern research stacks (e.g., JAX/Flax or PyTorch) and scaling experiments.

•
Strong experimental taste: Good judgment regarding baselines, ablations, and what is worth testing.

•
Comfort with scaling, evaluation methodologies, and diagnosing complex failure modes.

•
A focus on craft: You care about doing excellent work while maintaining a high velocity.

At Google DeepMind, we value diversity of experience, knowledge, backgrounds and perspectives and harness these qualities to create extraordinary impact. We are committed to equal employment opportunities regardless of sex, race, religion or belief, ethnic or national origin, disability, age, citizenship, marital, domestic or civil partnership status, sexual orientation, gender identity, pregnancy, or related condition (including breastfeeding) or any other basis as protected by applicable law. If you have a disability or additional need that requires accommodation, please do not hesitate to let us know.

Note: In the event your application is successful and an offer of employment is made to you, any offer of employment will be conditional on the results of a background check, performed by a third party acting on our behalf. For more information on how we handle your data, please see our Applicant and Candidate Privacy Policy.

Closing date: Tuesday, 17th March at 5:00pm GMT