Vincent Vanhoucke
Google DeepMind Robotics Head
Vincent Vanhoucke
Home > People > Vincent Vanhoucke
Profile
| Field | Details |
|---|
| Current Position | Google DeepMind Distinguished Scientist |
| Role | Google Robotics Head |
| Previous | Google Brain, Google Speech |
| PhD | Stanford University |
| Tenure | 20+ years at Google |
Key Contributions
- Google Robotics Head: Leading Google robot research including RT series, Gemini Robotics
- RT Series: Overall leadership of RT-1, RT-2, RT-X development
- Gemini Robotics: Leading Gemini 2.0-based robotics models
- Google Speech: Previously led speech recognition research
Research Timeline
Google Speech Era (2005-2015)
| Year | Work | Impact |
|---|
| 2005 | Joined Google | Google Speech team |
| 2012 | Deep Learning for Speech | Deep learning transition for speech recognition |
| 2015 | Google Now Voice Recognition | Commercial speech recognition improvements |
Google Brain & Robotics (2015-present)
Google Robotics Head
| Year | Work | Impact |
|---|
| 2016 | Started Robotics Research | Google Brain Robotics |
| 2018 | QT-Opt | Large-scale robot grasping |
| 2022 | RT-1 | Robotics Transformer |
| 2023 | RT-2 | First VLA model |
| 2023 | RT-X | Open X-Embodiment |
| 2025 | Gemini Robotics | Gemini 2.0-based VLA |
Leadership Role
Google Robotics Organization
- RT series research team leadership
- Everyday Robots (-> dissolved and integrated into DeepMind)
- Gemini Robotics team leadership
Key Decisions
- Set VLA direction (RT-2)
- Led Open X-Embodiment collaboration
- Gemini 2.0 and robotics integration
Key Projects Under Leadership
RT Series
| Model | Year | Significance |
|---|
| RT-1 | 2022 | Large-scale Robotics Transformer |
| RT-2 | 2023 | First VLA, Action as Language |
| RT-X | 2023 | 33 research lab collaboration |
Gemini Robotics
| Version | Features |
|---|
| Gemini Robotics | Gemini 2.0-based VLA |
| Gemini Robotics-ER | Embodied Reasoning |
| Gemini Robotics On-Device | Local execution |
Philosophy
Research/Development Philosophy
“The combination of large-scale data and powerful foundation models is the future of robot AI”
Google’s Robotics Strategy
- Foundation model utilization (Gemini)
- Large-scale data collection (Everyday Robots -> RT)
- Academic collaboration (Open X-Embodiment)
- Industry partnerships (Boston Dynamics, etc.)
Speech to Robotics Transition
Speech Recognition Experience
- Experienced importance of large-scale data
- Directly experienced impact of deep learning transition
- Experience in commercialization process
Application to Robotics
- Like speech, large-scale data is key for robotics
- Applied foundation model approach
- Utilized real deployment experience
Links
See Also