At Google DeepMind, the business problem is to advance the state of the art in artificial intelligence, specifically in building the next generation of conversational capabilities powered by the Gemini LLM, to empower multimodal conversational agents with groundbreaking speech and audio capabilities.
Requirements
- Significant industry experience building and deploying Speech / ML models.
- Demonstrated experience in data preparation, training, and evaluation of ML models.
- 5 years of experience with software development in Python, esp. ML frameworks like Tensorflow, JAX, PyTorch, etc.
- Experience with multimodal foundation models.
- Experience with real-time multimodal dialog systems.
- Hands-on experience with the Gemini models (data processing, training, SFT, RL, serving).
- Research background in NLP / Generative AI
Responsibilities
- Lead efforts to produce more natural and capable real-time dialog agents
- Collaborate closely with other teams on areas like reasoning, function calling, and multi-agent frameworks
- Partner with product teams design, develop, and deploy novel multimodal conversational agents.
- Create new data pipelines covering both real and synthetic sources, influencing pre-training and post-training (SFT & RL)
- Create real-time dialog capabilities for Gemini agents that seamlessly span audio, video, and text modalities.
- Pioneer end-to-end ML/Gemini architectures that streamline the dialog process, minimizing the need for complex model cascades.
- Direct impact on Gemini core model development - setting the direction for real-time dialog capabilities covering pre-training and post-training.
Other
- Bachelor degree in Computer Science, a related field, or equivalent practical experience.
- Ph.D. in Computer Science or a related field.
- Travel requirements not specified
- Clearance requirements not specified
- Diversity of experience, knowledge, backgrounds and perspectives and harness these qualities to create extraordinary impact.
