This site is a work in progress.

Ethan Keller

Ethan Keller

linkedin github email

I’m an AI researcher at Amazon AGI where I focus on LLMs for speech generation. Before that, I worked on reinforcement learning and federated learning at Imperial College London for sepsis treatment prediction.

My research interests include synthetic data, post-training, multimodality, interpretability, and automated research.

Journey

2024 - now
I joined the AGI team at Amazon to work on speech generation.
2022 - 2024

MSc Computing (AI & ML) at Imperial College London, where I worked with Aldo Faisal on federated reinforcement learning to predict treatments for sepsis patients.

After graduating, I joined the Brain & Behaviour Lab for about a year to work on the next generation of the AI Clinician.

2019 - 2022

BSc Computer Science and Engineering at the Delft University of Technology. My thesis focused on synthetic tabular data generation (FCT-GAN). I was head teaching assistant and taught 1000+ students in machine learning, data mining and programming courses.

Along the way I completed 2 internships. First at JetBrains Research (2021) researching lightweight messaging systems for robotics and IoT devices, and later at Amazon (2022) building web-based reading products in the Kindle team.

Languages

Proficient
English, French, Dutch
Intermediate
Hebrew, Spanish, German

Personal

Besides research, I like going to the gym, swimming, playing tennis and hanging out with friends. I also enjoy exploring topics in medicine and health or philosophizing about concepts like superintelligence, psychological biases, determinism or self-improvement (obviously without ever reaching a conclusion).