Leonardo D. Pepino
Teaching machines how to listen (and also speak)
š š š Iām a Research Scientist at Google Deepmind, working in bringing audio generation capabilities to Gemini.
I got my PhD in Computer Science š» at the Universidad de Buenos Aires š¦š· in 2025 and have worked at the intersection of deep learning š§ and audio š for more than 8 years.
My PhD research focused on self-supervised audio representation learning for general sound understanding and Iām particularly interested in:
- Representation Learning.
- Emotion recognition from speech.
- Neural audio codecs.
- Text to speech synthesis and voice cloning.
- Music source separation.
- Wakeword detection.
- NLP and textless NLP.
I also enjoy making music šø and exploring the outdoors through travel and hiking š šļø