Base Labs Research Scientist - RL Environments

Baseten•San Francisco, CA
•$210 - $285

About The Position

Base Labs, Baseten’s research lab, advances and democratizes open-source intelligence by publishing research, methods, and results for others to inspect, reproduce, and build on. Its training and inference infrastructure supports experiments at scale, while work with companies deploying models keeps its research connected to real-world needs. As a Research Scientist specializing in reinforcement learning environments, you will create open environments and study how models learn from them. You will design tasks and rewards, develop methods for generating and calibrating environments at scale, and run training experiments to understand their effects. The research includes studying how task diversity affects generalization, how curricula should adapt as models improve, and which reward structures support reliable learning. Findings will inform released environments and the lab’s broader training research. The role is open across levels, with scope matched to experience. You will work closely with research engineers, applied researchers, and the wider Base Labs team. Senior hires will help establish research directions and mentor others. All team members contribute to the environments, methods, and published evidence supporting the program.

Requirements

  • Evidence of being able to choose a useful research question, design an informative experiment, implement it, and reason carefully about the results.
  • Depth in a relevant area such as reinforcement learning, model evaluation, agent systems, synthetic data, or experimental methodology.
  • Ability to write code to build environments, run experiments, and inspect failures in detail.

Nice To Haves

  • Experience with LLM training.
  • Experience building something that other researchers use.
  • Evidence of research ability through a paper, open-source project, or strong experimental investigation.
  • A PhD or previous Research Scientist title is not required; exceptional candidates from adjacent fields are welcome.

Responsibilities

  • Identify capabilities and failure modes worth studying, informed by real model use, including areas such as software development, tool use, scientific reasoning, and following complex instructions.
  • Design task distributions, interactions, rewards, and verifiers together; establish that tasks are solvable and that success depends on the intended capability; investigate shortcuts, ambiguous instructions, and grading weaknesses.
  • Build scalable methods for environment creation while maintaining quality; study ways to generate diverse tasks, calibrate difficulty, and detect defects before they spread through a large training pool.
  • Run reinforcement learning experiments to measure learning and transfer; compare curricula, task mixtures, and reward designs; investigate failed runs and unexpected behavior; and test whether improvements hold on unseen tasks.
  • Develop evaluation methods that distinguish capability differences from sampling noise, grader errors, or artifacts of agent tools and execution setups, and preserve evidence needed to reproduce and challenge conclusions.
  • Work with safety researchers on environments that reveal reward hacking and misalignment, and test ways to teach more reliable behavior.
  • Release environments, datasets, training recipes, and research findings for use by other teams, and clearly describe their assumptions, limitations, and measured effects.

Benefits

  • Competitive compensation, including meaningful equity.
  • For U.S. employees and dependents: 100% coverage of medical, dental, and vision insurance.
  • Flexible PTO policy.
  • Paid parental leave.
  • Fertility and family-building stipend through Carrot.
  • For U.S. employees: company-facilitated 401(k).
  • Exposure to a variety of ML startups, offering learning and networking opportunities.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service