Early employee at Thinking Machines Labs.

7 years at Apple

Low-Latency Cascaded Conversational Agent in MLX
Led the engineering effort to build a cutting edge conversation agent using novel research in ASR, LLMs, TTS, VAD, and Vocoders.

Federated Evaluation and Tuning for On-Device Personalization: System Design & Applications
Framework for distributed training of models on iPhones, Macs, and other Apple products in a privacy preserving manner