Post-Training Researcher (AI)

Location: San Francisco, United States; on-site

Company

Building Devin, an AI software engineer, and advancing AI systems that reason about real-world tasks.

What you will do

• Develop and iterate on post-training recipes covering datasets, training stages, and hyperparameters.
• Design evaluations that measure meaningful model and production performance while maintaining evaluation integrity.
• Investigate unexpected training results and identify the underlying causes of model behavior.
• Apply and advance RLHF, RLAIF, preference modeling, reward learning, and constitutional approaches for agent alignment.
• Measure scaling behavior across data and compute, and develop new methodologies when existing approaches reach their limits.
• Work across research and engineering to move prototypes into real deployment.

Requirements

• Demonstrated work advancing ML systems through post-training, alignment, or related methods.
• Strong foundations in probability, statistics, and machine learning theory.
• Original contributions through publications, open-source work, or equivalent industry results.
• Experience with large-scale distributed training and debugging.
• Systems-level understanding of the interaction between training pipelines, data, and evaluation.
• Ability to work effectively in ambiguous, fast-moving research environments.

Nice to have

• PhD or another credential demonstrating research capability; credentials are considered alongside demonstrated ability.

Culture & Benefits

• Small, highly selective team where research and product development move together.
• Rapid path from prototypes to real deployment.
• Large compute allocations with training jobs running across thousands of GPUs.
• Environment emphasizing speed, autonomy, technical depth, and minimal process overhead.
• Resources for operating at frontier scale from the start.