Nal Kalchbrenner

Nal Kalchbrenner is a deep-learning researcher known for foundational work in generative and sequence models. After a PhD at the University of Oxford, he spent years at Google DeepMind, where he helped establish the Google Brain team in Amsterdam. He was a key contributor to several landmark architectures, including WaveNet — the raw-audio generative model that reset the state of the art in text-to-speech and now powers synthetic voices used by hundreds of millions of people — as well as ByteNet, an early linear-time approach to neural machine translation, and the PixelRNN/PixelCNN family of image models.

His research spans natural-language understanding, image and video modelling, speech and audio synthesis, and reinforcement learning — a breadth that places him among the researchers who defined the generative-AI toolkit before it reached the mainstream. WaveNet's generative approach was subsequently adopted across the speech industry, including in widely deployed voice assistants.

Kalchbrenner is also active as an early-stage investor in AI and deep-tech companies. Dealroom tracks his participation in the early financing of Perplexity, the AI answer engine, placing him among the cohort of research leaders who backed the company before it scaled. His investing reflects the same generative-modelling expertise that shaped his research career.