Juan Manuel Coria is Co-Founder and CTO at pyannoteAI, a speaker intelligence platform that enables AI to identify who is speaking in audio recordings across any language.

Juan earned his PhD in Artificial Intelligence from Université Paris-Saclay, focusing on Continual Representation Learning in neural networks for speech and language, and co-authored the open-source pyannote.audio toolkit used by over 100,000 developers.

Juan published multiple research papers on speaker diarization and metric learning, including work on overlap-aware low-latency speaker diarization and real-time audio processing, and created the diart toolkit for AI-powered real-time audio applications.

pyannoteAI raised $9M in seed funding led by Crane Venture Partners and Serena in 2025, with backing from HuggingFace CTO Julien Chaumond and former Meta/OpenAI researcher Alexis Conneau.

The pyannote.audio framework achieves over 45 million monthly downloads on HuggingFace and ranks among the platform's top 20 most downloaded models, powering voice AI applications for enterprises worldwide.