Mistral researcher to keynote Interspeech 2026 on Voxtral audio foundation models

shrutirij · x · 2026-09-22

Khyathi Chandu of Mistral AI announced she will give a keynote at Interspeech 2026 on Oct 1, covering Mistral's audio foundation models Voxtral and Voxtral TTS — giving an LLM ears and a voice.

She is also involved in the PTSFM special session on post-training of speech foundation models, which focuses on adaptation, alignment, reasoning, and efficiency (compression/distillation/continual learning), plus persistent challenges like unbalanced data and evaluating alignment.

Original post →

More from Companies & People

Companies & People channel →