On garden leave since July 9th 2026.

Generative Media team / Gen AI. Lyria 3 post-training and evals lead; cross-modal generation in Veo/Gemini Omni; artist-centric tools.

2026:

• Lyria 3 launched in Gemini on Feb 18th

• Lyria 3 launched in Flow Music on Feb 24th

• Lyria 3 Pro launched in Gemini on Mar 25th

• Lyria 3 Pro and Clip launched in Vertex AI, Gemini API, Google AI Studio, Google Vids, Flow Music on Mar 25th

• Gemini Omni launched at Google I/O

• Lyria 3.5 launched in Flow Music on July 29th

Deep Learning, then Generative Media team.

• Delivered live demos of our music AI tech to industry stakeholders (artists, managers, labels) and at various venues (closed-door events, press demos, conferences).

• Leadership: Co-lead of GenMusic—team of ~40 whose work was presented at impactful events such as Google I/O and is leading the responsible development of music AI, in partnership with artists, producers and music creatives. GDM tech lead for the Youtube Shorts Dream Track quality workstream—coordinated with several teams to drive model progress, hit the quality launch bar and inform leadership product decisions.

• Individual technical contributions: Lyria and Music AI Tools (Sandbox)—model capabilities and finetuning for product use cases, identifying signals in large-scale datasets, designing and running human evals.

Deep Learning team.

• Leadership: Co-led internal projects on multimodal learning and audio / music generation.

• Individual technical contributions: Conducted research on multimodal and long-range generative methods; published at ICML.

• Mentoring: Hosted a research intern whose project got published in TMLR. Delivered a tutorial at EEML and mentored 5 students for a research project at LOGML.

Master’s research projects: Structure-aware Generation of Molecules in Protein Pockets (Pavol Drotar, 2020-21) (⁹²⁄₁₀₀) (presented at NeurIPS MLSB), Machine Unlearning (Mukul Rathi, 2020-21) (⁹¹⁄₁₀₀), Goal-Conditioned Reinforcement Learning in the Presence of an Adversary (Carlos Purves, 2019-20) (⁸⁷⁄₁₀₀), Representation Learning for Spatio-Temporal Graphs (Felix Opolka, 2018-19) (⁸⁵⁄₁₀₀) (presented at ICLR RLGM), Dynamic Temporal Analysis for Graph Structured Data (Aaron Solomon, 2018-19) (presented at ICLR RLGM)

Computer Science Tripos Part II projects: Benchmarking Graph Neural Networks using Wikipedia (Péter Mernyei, 2019-20, Novel Applications spotlight talk at ICML GRL+), Multimodal Relational Reasoning for Visual Question Answering (Aaron Tjandra, 2019-20), The PlayStation Reinforcement Learning Environment (Carlos Purves, 2018-19) (⁸⁰⁄₁₀₀) (presented at NeurIPS Deep RL), Deep Learning for Music Recommendation (Andrew Wells, 2017-18) (⁷⁶⁄₁₀₀).

Undergraduate courses for Murray Edwards, King’s, and Queens’ Colleges: AI, Databases, Discrete Mathematics, Foundations of Computer Science, Logic and Proof, Machine Learning and Real-world Data.