Base Labs Research Scientist - RL Environments
Baseten is hiring a Base Labs Research Scientist - RL Environments for a remote role open to applicants in United States. Level rates it ; you can apply on Baseten.
Level's read on this job
- Level rates this research scientist role Builds AI ●●●●: the job is building AI systems.
- This is listed as a remote role.
- The posting mentions Cursor, Clay and AI evaluation among its tools and skills.
- Baseten has 103 other open roles on Level.
- Posted in the last few days.
AI in this role
ABOUT BASETEN
Baseten powers mission-critical inference for the world's most dynamic AI companies, like Cursor, Notion, OpenEvidence, Abridge, Clay, Gamma, and Writer. By uniting applied AI research, flexible infrastructure, and seamless developer tooling, we enable companies operating at the frontier of AI to bring cutting-edge models into production. We're growing quickly and recently raised our $1.5B Series F, led by Altimeter Capital, Conviction Partners, and Spark Capital. Join us and help build the platform engineers turn to ship AI products.
ABOUT BASE LABS
Base Labs is Baseten’s research lab, working to advance and democratize open-source intelligence. We publish our research, methods, and results so others can inspect, reproduce, and build on them. Baseten’s training and inference infrastructure gives us a place to test ideas at scale, while our work with companies deploying models keeps us close to the problems those models need to solve.
THE ROLE
Models learn from the environments we give them. If a reward is easier to exploit than a task is to solve, a model can learn the shortcut. If a task measures the wrong capability, a higher score tells us little about progress. Building good environments requires decisions about what to teach, how to measure it, and how to establish whether it transfers.
As a Research Scientist at Base Labs specializing in RL 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 the training experiments needed to understand their effects.
There is substantial reinforcement learning research in this work. We want to understand how task diversity affects generalization, how curricula should change as models improve, and which reward structures produce reliable learning. Your findings will inform both the environments we release and the broader lab’s training research.
RESPONSIBILITIES
Identify capabilities and failure modes worth studying, informed by real model use. These may include 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 weaknesses in grading.
Build methods that scale environment creation without weakening the quality standard. Study how to generate diverse tasks, calibrate difficulty, and detect defects before they spread through a large training pool.
Run RL experiments to measure learning and transfer. Compare curricula, task mixtures, and reward designs; investigate failed runs and unexpected behavior; test whether improvements hold on unseen tasks.
Develop evaluation methods that distinguish capability differences from sampling noise, grader errors, or artifacts of the agent’s tools and execution setup. Preserve the evidence needed to reproduce and challenge your 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 that other teams can use. Make their assumptions, limitations, and measured effects clear.
WHO WE'RE LOOKING FOR
We look for evidence that you can choose a useful research question, design an informative experiment, implement it, and reason carefully about the result.
You should bring depth in a relevant area such as reinforcement learning, model evaluation, agent systems, synthetic data, or experimental methodology. You should be comfortable writing the code needed to build environments, run experiments, and inspect failures in detail. Experience with LLM training is valuable, as is experience building something that other researchers use.
Your evidence might be a paper, an open-source project or a strong experimental investigation. We also welcome exceptional people from adjacent fields. A PhD or a previous Research Scientist title is not required.
WORKING HERE
We are hiring 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. Everyone will contribute to the environments, methods, and published evidence on which the program grows.
BENEFITS
Competitive compensation, including meaningful equity
(U.S. only) 100% coverage of medical, dental, and vision insurance for employee and dependents
Flexible PTO policy
Paid parental leave
Fertility and family-building stipend through Carrot
(U.S. only) Company-facilitated 401(k)
Exposure to a variety of ML startups, offering unparalleled learning and networking opportunities.
Apply now to embark on a rewarding journey in shaping the future of AI! If you are a motivated individual with a passion for machine learning and a desire to be part of a collaborative and forward-thinking team, we would love to hear from you.
At Baseten, we are committed to fostering a diverse and inclusive workplace. We provide equal employment opportunities to all employees and applicants without regard to race, color, religion, gender, sexual orientation, gender identity or expression, national origin, age, genetic information, disability, or veteran status.
We are an Equal Opportunity Employer and will consider qualified applicants with criminal histories in a manner consistent with applicable law (by example, the requirements of the San Francisco Fair Chance Ordinance, where applicable).
How we rate this
Base Labs Research Scientist - RL Environments at Baseten rates 96 out of 100 for how much of the daily work is AI. That makes it Builds AI (AI Level 4 of 4). The level is about AI in the job, not seniority.
Builds AI. The job is building AI systems.
- ●●●● Builds AI80 to 100
- ●●●○ Works on AI60 to 79
- ●●○○ Uses AI40 to 59
- ●○○○ Little AI0 to 39
Levels come from how often the tools, models and workflows of the role are named in the posting itself. Read the description and count the AI tools, models and workflows it names.
Prepare for this job
Built from this posting and from similar jobs on Level.
Skills and AI tools this role asks for
Also common in research scientist jobs at this level
Pay
- Typical research scientist job
- $190k
Use the typical research scientist figure as a reference when pay comes up. See research scientist jobs
Questions you could be asked
- How do you decide that one model's output is better than another's for a given task?
- Tell me about a research question you investigated. What did you find?
- What are the limits of Cursor that you've run into, and how did you work around them?
- What's a project where you used Clay hands-on?
- How would you decide a model or AI system is ready to ship?
Adapt your resume
- List these exact terms on your resume: AI Evaluation, AI Research, Cursor, and Clay. An applicant tracking system matches the wording, not the idea.
- Attach one line of real, concrete experience to at least one of them. A tool named with nothing behind it rarely survives a human read.
- Lead with what you built, trained or shipped. This role is judged on the AI system itself, not the tools around it.
Get new remote research scientist jobs (Builds AI ●●●●) by email
One email a week, each job rated for how much AI is in the work. No recruiter spam, unsubscribe in one click.
Free.
Similar roles
Research roles that build AI, at other companies.
What kind of AI work fits you?
Answer 12 practical questions in about three minutes. Get a simple profile, the work it points to, and live roles to explore next.
Find my next step