Chris Cundy

Chris Cundy

Machine Learning PhD Student

Stanford University


I am broadly interested in Artificial Intelligence (AI), particularly in ensuring that sophisticated AI systems will robustly and reliably carry out the tasks that we want them to.

If you’re a student at Stanford (undergraduate/masters/PhD) who wants to work on a project involving safe and reliable machine learning: get in touch!

I studied Physics as an undergraduate at Cambridge University before switching area to take a computer science master’s degree. During my master’s, it was a pleasure to work with Carl E. Rasmussen, developing variational methods for Gaussian Process State-Space Models.

Before starting my PhD at Stanford, I worked at the Centre for Human Compatible AI, collaborating with Stuart Russell and Daniel Filan. I have also worked at the Future of Humanity Institute at Oxford University, collaborating with Owain Evans on scalable human supervision of complex AI tasks.

Get in touch at chris dot j dot cundy at gmail dot com


  • Probabilistic Machine Learning
  • Generative Models
  • Reinforcement Learning
  • Safe and Reliable ML


  • PhD in Computer Science, 2018-Ongoing

    Stanford University

  • MEng in Computer Science, 2017

    Cambridge University

  • BA in Natural Sciences (Physics), 2016

    Cambridge University

Recent Publications

Quickly discover relevant content by filtering publications.

Privacy-Constrained Policies via Mutual Information Regularized Policy Gradients

As reinforcement learning techniques are increasingly applied to real-world decision problems, attention has turned to how these …

Recent Posts

Installing cdt R prerequisites in Ubuntu without root

I wanted to use some of the tools from the causal discovery toolbox which require R and the pcalg package to be installed. As a complete newcomer to R, it was more hassle than I thought it would be to install R on an ubuntu server without root access.

Managing ArXiv RSS Feeds in Emacs

Background It’s very important for any researcher to keep up with the papers that are being published, especially in the fast-moving field of machine learning. However, there are a lot of papers from the arxiv categories which I follow, sometimes hundreds of papers a day.

Mutual Information Regularization for Reward Hacking

Introduction In some recent work that we just put up on arxiv, we explore the idea of training reinforcement agents which obey privacy constraints. Something I’ve wanted to explore for a while is the possibility of using this constrained RL approach for dealing with reward hacking.