Audio to Audio Research Scientist, DeepMind

Job description

Develop audio-first models capable of orchestrating and planning complex dialogs, including leveraging external tools like search when necessary. Leverage new sources of data (real and synthetic) to empower new real-time dialog capabilities. Work with infra teams to design models suitable for open-mic conversations that are fluid and have low latency. Create comprehensive datasets and benchmarks to improve and measure audio-to-audio (A2A) model performance on conversationality. Minimum Qualifications: PhD degree in Computer Science, Electrical Engineering, or Computer Engineering, a related field, or equivalent practical experience. Experience in artificial intelligence or machine learning. Experience with large language models, NLP, or Generative AI. Experience in an applied research setting. Experience working with machine learning (ML) models, training methodologies, or evaluation datasets. Preferred Qualifications: Experience with Generative Audio Modeling, Multimodal modeling, On-device AI Implementation or Fine tuning for LLMs. Experience in NLP, LLMs or Generative AI. A strong publication record in conferences or journals in the field of communications or signal processing.