My name is Timo, I am a Research Engineer based in the San Francisco Bay Area. At Google DeepMind I work with the Gemini team on post-training for audio.
In my leisure time I like to fly little airplanes, 3d print, and occasionally work on artsy projects. Sports-wise road cycling and hiking are my favorite activities.
To get a deeper idea of who I am you are invited to take a look at my values. An overview of recent projects is given at blog.timodenk.com.
I am occasionally publishing articles at blog.timodenk.com.
Here are some selected ones:
(for a full list see my Google Scholar profile)
Timo Denk, Yu Takagi et al. (Jul 2023): Brain2Music: Reconstructing Music from Human Brain Activity. [paper] [samples]
Andrea Agostinelli, Timo Denk et al. (Jan 2023): MusicLM: Generating Music From Text. [paper] [samples] [data]
Timo Denk, Ana Peleteiro Ramallo (Dec 2020): Contextual BERT: Conditioning the Language Model Using a Global State. Workshop paper accepted at the COLING 2020 workshop TextGraphs-14. [paper] [presentation]
Timo Denk, Christian Reisswig (Sep 2019): BERTgrid: Contextualized Embedding for 2D Document Representation and Understanding. Best Paper Award at the NeurIPS 2019 Document Intelligence workshop. [paper] [presentation] [thesis]
