Research Scientist - Post Training

ABOUT THE ROLE

The company is building out its post-training research team and hiring 2–3 Research Scientists to work together.

The job is to prove that the company's data works. You'll design and run controlled training experiments that isolate how its datasets affect model behavior.

Using SFT and RL-based post-training, you'll measure how different data sources affect

capability, generalization and alignment. You'll then turn those results into practical

recommendations for how the data is structured, selected and produced.

Working closely with partner AI labs, you'll turn the datasets into clear, defensible evidence: this data leads to this improvement under these conditions. You'll also help build shared

infrastructure and benchmarks that make these findings reproducible across models,

and benchmarks that make these findings reproducible across models, capabilities and domains. It'sexperimental, customer-facing, high-impact work at the edge of model development.

RESPONSIBILITIES

Run controlled SFT and RL experiments to measure how the company's datasets affect model performance.

Help build public evals and new data types that push the frontier

Publish external-facing research, blog posts and technical reports

Work with internal project leads to improve data quality based on your results

REQUIRED QUALIFICATIONS

Must have experience at a frontier AI lab

Strong familiarity with LLM training and evaluation methods

Able to design lightweight experiments, move fast and pull actionable insights from messy results

Comfortable working across domains, including finance, software engineering and policy

A bias toward building over theorizing

BENEFITS AND PERKS

Profit sharing (total cash compensation of $250,000 – $450,000)

Health insurance

Free gym membership

Daily lunch and dinner covered

Pay: $150,000.00 - $250,000.00 per year

Benefits:

• Employee stock purchase plan
• Gym membership
• Health insurance
• Stock options

Work Location: In person