Organize and optimize post-training initiatives across hardware, physical systems, and infrastructure into clear goals, milestones, and release plans. Scope and execute domain-expert data acquisition, design ingestion pipelines, synthetic task generation, and RL environments. Partner with modeling leads to build customer-centric benchmarks and simulation/rule verifiers grading executability, physical well-formedness, and spec satisfaction.
Manage project planning using lightweight methodologies to improve research velocity, release predictability, and align cross-functional dependencies. Evaluate technical trade-offs between geometric plausibility versus true functional engineering correctness, manufacturability constraints, and reward hacking risks.
Job description provided by Google.
View the original posting