Research Scientist - Environments, Data and Post-Training
At a glance
Mid-level AI Research role at Mercor. San Francisco · full-time · $250,000–$500,000 base.
aimlroles summary, based on the employer's posting.
What you'll do
- Build post-training approaches that strengthen model reasoning, tool use and agent behavior
- Compare training data, reward setups, environments and optimization choices through controlled experiments
- Create large-scale reinforcement-learning pipelines, including verifiable-reward training
- Assess model strengths and breakdowns, then design focused interventions and data-quality measures
- Develop data pipelines, evaluation rubrics and benchmarks, and share findings through research outputs
What you bring
- You have trained and evaluated machine-learning models in practice
- You bring a strong research record in post-training, reinforcement learning, language-model evaluation or a related area
- You can assess model behavior, experimental findings and data quality with rigor
- You program strongly and have implemented machine-learning systems
Who this fits
This role suits a research scientist who can turn open research questions into experiments and working systems, and collaborate with researchers, engineers and domain experts. You’ll work in person in San Francisco five days a week, with remote Saturdays optional.
From the employer
About Mercor
Mercor's mission is to organize human intelligence to power the AI economy. We're a leading AI data company, building the layer between human expertise and frontier models. Millions of domain experts on the platform are paid over $4 million per day to train frontier AI models. Mercor's APEX benchmark family measures AI's real-world impact on professional work. Mercor Enterprise brings this same infrastructure to Fortune 500 companies: helping companies capture how their best people actually work, translating that expertise directly back into agents.
Mercor is creating a new category of work where expertise powers AI advancement. Achieving this requires an ambitious, fast-paced and deeply committed team. You’ll work alongside researchers, operators, and AI companies at the forefront of shaping the systems that are redefining society. Mercor is a profitable Series C company valued at $10 billion. We work in-person five days a week in our San Francisco, NYC, or London offices.
About the Role
As a Research Scientist at Mercor, you will work at the intersection of research and engineering on frontier post-training. You will develop new training and evaluation methods, test them through rigorous experiments, and implement successful approaches at scale.
Working with researchers, engineers, domain experts, and customers, you will investigate how data, rewards, environments, and optimization methods shape model behavior. Your work will influence frontier models, Mercor’s products, and the broader research community through releasing blogposts, technical reports, and papers.
What You’ll Do
Implement novel post-training methods that improve model reasoning, tool use, and agentic behavior.
Develop new training recipes for frontier open models.
Design and run experiments across datasets, reward functions, environments, and optimization strategies, including methods such as GRPO and DAPO.
Build reinforcement learning with verifiable rewards (RLVR) and other post-training pipelines at scale.
Investigate model capabilities and failure modes, then develop targeted training interventions.
Create methods for measuring data quality, usability, and causal impact on model performance.
Build scalable pipelines for data generation, filtering, augmentation, and selection.
Develop rubrics, evaluators, benchmarks, and scoring systems that inform training decisions.
Translate open-ended research questions into rigorous experiments and production systems.
Collaborate with researchers, applied AI teams, engineers, and domain experts producing training data.
Contribute to open-source post-training tools and research.
What We’re Looking For
Demonstrated experience training and evaluating machine learning models.
A strong research record in post-training, reinforcement learning, language-model evaluation, data-centric ML, or a closely related field.
Ability to reason rigorously about model behavior, experimental results, and data quality.
Strong programming skills and experience implementing machine learning systems.
Knowledge of the current AI research landscape and important open problems.
Excitement to work in person in San Francisco, five days a week (with optional remote Saturdays), and thrive in a high-intensity, high-ownership environment.
Nice To Have
Experience on an industry post-training or frontier-model team.
Main authorship of publications at top-tier conferences (NeurIPS, ICML, ACL).
Experience with synthetic-data generation
Experience building large-scale evaluation or data-generation infrastructure.
Solid foundations in distributed or backend systems, and experimental design.
Familiarity with APIs, databases, and cloud infrastructure.
Benefits
Bi-annual performance bonus structure
Generous equity grant vested over 4 years
Up to $15k Relocation bonus
$10K housing bonus (if you live within 0.5 miles of our office)
$1.5K monthly stipend for meals
Free Equinox membership
$200 monthly laundry reimbursement
$200 monthly personal wellness reimbursement
Health, Dental, Vision insurance