We think bootcamps are among the best ways to quickly get up to speed in deep learning and develop skills useful for conducting AI safety research

We are planning to run a 3-week bootcamp in January, 2026 following the ARENA curriculum.

The bootcamp will be in-person in our Harvard Square office, with dedicated teaching assistants. Prerequisites are familiarity with basic Python, linear algebra, and calculus. By the end of the program, participants will be reproducing key interpretability papers and will have hands-on experience with Transformers, reinforcement learning, and evals.

If you’re interested in participating, please fill out our expression of interest form here!

We expect to open applications in early November.