Job Summary
The position focuses on creating methodologies for data quality and labeling, aimed at embedding both physical and action-based knowledge into Gemini foundation models. The role involves designing and implementing training ablations and developing research tools to analyze and validate the performance of robot policies.
Responsibilities
- Develop methodologies for enhancing data quality and labeling for Gemini foundation models.
- Integrate physical and action knowledge into the modeling process.
- Design and execute training ablation experiments.
- Construct research tools for analyzing and verifying the performance of robot policies.
Qualifications
- PhD in Computer Science or a related technical discipline.
- Experience in robot policy learning and research publication.
- Hands-on experience with simulators and robotic platforms.
- Familiarity with systematic benchmarking techniques for foundation models.
