Research Engineer, Post-Training

Join an applied AI lab as a Research Scientist focused on post-training to enhance the capabilities, safety, and effectiveness of AI software agents like Devin. This role involves iterating on training recipes, developing robust evaluations, and advancing alignment techniques to shape agent behavior and performance.

• Develop and iterate on post-training recipes including datasets, training stages, and hyperparameters
• Design and implement evaluations to measure model behavior and production performance
• Investigate and understand the root causes of unexpected training results
• Apply and advance alignment techniques like RLHF and RLAIF for agent behavior
• Analyze performance scaling with data and compute, developing new methodologies
• Contribute to a fast-paced research environment with significant compute resources