About
Building agentic AI systems at Yutori. Previously led multimodal post-training of Llama for AI Studio at Meta, enhancing multimodal and tool-use capabilities while also improving model safety. Founding member of Google's "Look and Talk" feature, launched at Google I/O 2022 — led pretraining video-audio models, post-training the multimodal Gemini model, optimizing inference, and large-scale data collection.
Badges
Maker History
🎉
Joined Product HuntFebruary 19th, 2024