Requirements

• Experience building systems powered by LLMs or other ML models.
• Experience in software development in training, developing or applying machine learning models.

Nice to haves

• Strong AI development knowledge, training GenAI models.
• Strong software engineering fundamentals, with a track record of shipping production-quality code.
• Strong collaboration and communication skills across research, engineering and product teams.
• Interest in AI agents and how they can make people more productive.
• Comfortable working in an ambiguous, changing research and product environment.

What you'll be doing

• Build and optimize multi-agent hierarchical pipelines and cross-application reasoning agents for productivity agents.
• Develop, scale, and maintain automated evaluation raters to assess model capabilities under complex, real-world conditions.
• Maintain secure execution environments (Python Workspace API harnesses, Bash/gws CLI) to evaluate agent trajectories hermetically, and allow for large-scale distillation and training in the environments.