Research Engineer, Post-Training

Highlights

• Shape how AI agents learn from raw capability to real-world usefulness.
• Iterate training recipes, evals, and alignment for Devin and future systems.
• Scale methods with data and compute when existing approaches hit ceilings.

What they offer

• Large compute allocations with training across thousands of GPUs
• Small selective team where prototypes reach real deployment quickly
• Minimal process with autonomy, speed and technical depth

Ideal candidate

ML researcher-engineer obsessed with why training behaves as it does, with a track record advancing models via post-training and alignment. Thrives in ambiguous, fast-moving environments and cares about shipping agents that work in the real world.

Requirements

• Track record advancing ML systems through post-training, alignment, or related methods
• Strong fundamentals in probability, statistics, and ML theory
• Ability to distinguish real effects from noise and bugs in experimental data
• Evidence of original contributions via publications, open-source impact, or industry results
• Experience with large-scale distributed training and debugging
• Systems-level thinking across training pipelines, data, and evaluation
• Comfort with ambiguity and fast-moving research environments