Research / Data Engineer

About the Company

An early-stage AI company, founded in 2025, that builds long-horizon reinforcement learning environments and post-training data for frontier AI labs. Its customers include the leading US frontier labs as well as top international labs, with active engagements across them. Unlike most data vendors, which treat post-training data as a human-services problem, the company treats it as a technical and research problem. The long-term vision is a model that can produce post-training data on its own, so output scales with talent and not with headcount. The company is cash-flow positive, recently raised a seed round, and has a small, elite team with backgrounds in published research, distributed systems, and top AI and technology companies. It recently released a benchmark that drew strong interest from frontier labs.

About the Role

Inbound demand for post-training data quality is higher than the company can currently fulfill, so this hire is about scaling the talent to meet it. The shape that works best is a domain or research expert with a strong ML or engineering spike: someone who both does research and produces data, because in this company's thesis the two are inseparable. For this early hire, the research and data-production archetypes are deliberately merged into one person. The technical scope is broad, from provisioning infrastructure to curating high-quality post-training data, and a core skill is data taste: good judgment about what problems pervade a domain and what makes a realistic, economically valuable task or environment that a lab will actually buy. The team screens first for research and ML horsepower, and above all for high slope.

What You'll Own

• Post-training models end to end, from infrastructure provisioning through curating high-quality post-training data
• Producing RL environments and long-horizon tasks across domains
• Scaling data generation superlinearly through self-improving developer processes, not more people
• Developing data taste: realistic, economically valuable tasks and environments that labs will buy
• Blending research and data production as one function
  
  

What We're Looking For

• A genuine spike in ML or AI research (post-training, RL, evals, interpretability, or similar) combined with strong software engineering. The dual competency is the whole shape of the role.
• The ability to post-train models and to build and curate high-quality post-training data
• Data taste, or the research horsepower to develop it fast
• Fast code comprehension and problem-solving, including debugging ML pipelines in PyTorch
• Domain or research expertise in a specific field alongside an ML and engineering core. Neuroscience, physics, chemistry, and systems backgrounds all fit.
• Organic, obsessive interest shown through publications, side projects, open-source work, or writing
• A mindset that treats data as a research problem, not a headcount problem
• High slope, deeply committed, and results-driven
• Excitement about the superlinear-scaling thesis and about working with frontier-lab customers
• 3 to 7 years of experience
  
  

Logistics

San Francisco, in person, Compensation is $250K to $300K base plus equity. Visa sponsorship is available.