The posting
ABOUT
Sampura Research is an AI safety research non-profit focused on scalable oversight for frontier models through improved human+AI “judges”.
Judges (LLM, human, or a hybrid) are essential for aligning highly capable models through their use in evaluation and training. Limitations in today's judges can lead to undesirable behavior such as reward hacking in training, or failure to spot misalignment in evaluation. As models become more capable and their data more complex, often beyond what the systems overseeing them can understand, we expect these limitations to grow in severity and to become harder to notice.
We are designing novel judge protocols which leverage the strengths of both humans and models to outperform either alone, and building high-quality evaluations to explore judge performance across a diverse set of safety-related tasks. We believe that our focused approach will rapidly advance the science of judges by providing a much-needed systematic measurement of which protocols actually work, and producing state-of-the-art judges which can be used in real world training and evaluation pipelines.
Backed by an $11M grant from Coefficient Giving ($7M for the first year, and another $4M pledged), Sampura Research was founded in 2026 by members of DeepMind’s alignment and human data teams.
Please see our announcement https://sampura.org/news/announcing-sampura-research/ for more information.
ROLE OVERVIEW
As a founding Member of Technical Staff, you will help shape the research and engineering direction alongside the founding team, and own end-to-end research and engineering problems within our agenda:
Evaluation: Build diverse and high-quality evaluations for judge protocols, spanning static datasets (existing and novel) and dynamic settings where policies actively optimize against the judge during RL
Methods: Research and implement strategies for improving human+AI judge protocols, such as learning-to-defer, AI-generated assistance for human raters, and co-training LLM judges alongside the policy during RL.
Infrastructure: Build the tools required to enable and scale our research agenda, including large-scale evaluation and training pipelines, a human rating platform, and a public judge leaderboard.
We've initially budgeted ~$1M per quarter for compute and human data, across a team of fewer than ten, to ensure that we can commit to ambitious goals in all of these areas. You'll scope ambiguous problems independently, communicate progress clearly, and synthesize findings into papers, blog posts, and benchmark releases.
Staff hires may also lead research and mentorship programs through paid fellowships or volunteer research focused on evaluation or methods work.
If our work and mission resonate with you, please consider applying even if you don’t tick every box. We strongly believe in investing in excellent people regardless of prior or formal experience.
You may be a good fit if you have
- A Bachelor's degree or higher in Computer Science, Mathematics, or a related field or equivalent experience.
- Strong Python engineering skills and experience in large collaborative codebases, with experience in the modern ML stack (e.g., LLM inference APIs, PyTorch/JAX, data analysis tooling) and large codebases.
- Experience training and evaluating large language models, including critically assessing datasets and analysing results quantitatively and qualitatively.
- A willingness to embrace AI tools intelligently, applying scrutiny and human review where necessary.
- A track record of exploring and resolving open-ended research questions with rigor and integrity, in academic or industry labs, fellowships, or volunteer research programs.
Outstanding candidates will have
- Demonstrated ownership across multiple parts of the AI research lifecycle (experiment design and execution, infrastructure, data collection and processing, inference).
- Expertise in a highly relevant domain such as model confidence/calibration, interpretability, safety/alignment evaluation, human-computer interaction, or other fields related to scalable oversight.
- Hands-on experience with post-training methods such as reinforcement learning, adversarial fine-tuning, or distillation, especially applied to judges or reward models.
- Hands-on experience with human data collection and rater study design.
- Leadership as a mentor or team lead for related research, either inside a research lab or via programs like SPAR, MARS, or MATS.
SALARY RANGE:
Junior hires: £100,000-£150,000 (~$136,000-$205,000)
Senior hires: £150,000-£200,000 (~$205,000-$273,000)
Staff hires: £200,000-£290,000 (~$273,000-$395,000)
*Approximate USD conversion as of 24 August 2026.
Exceptional students/new-grads and early-career individuals will often come in at the junior level. Researchers or engineers with several relevant years of experience will often come in at the senior level. Staff hires typically bring deep expertise in a directly relevant area at notable research labs, and have led research programs or teams.
We'll agree a target level with you early in the process, and may reassess as you progress.
LOGISTICS:
Location: This role is based on-site in London.
Visa Sponsorship: We offer visa sponsorship for this role.
Start Dates: We are hiring 2-4 FTEs targeting an October 2026 start, and hiring other FTEs on a rolling basis.
Deadlines: We will review applications on a rolling basis, so please apply ASAP. For October hires, apply by Sept 21 latest. Feel free to apply again in a few months, if your experience has updated.
EQUAL OPPORTUNITY:
Sampura Research is an equal opportunity employer. We are committed to building a team with diverse backgrounds and perspectives and believe this makes our research and workplace better. We do not discriminate on the basis of race, religion or belief, sex, sexual orientation, gender reassignment, age, disability, marriage or civil partnership, pregnancy or maternity, or any other protected characteristic. If you need adjustments at any stage of the application process, let us know and we'll make them.
