Our Team

  • Callum McDougall

    FOUNDER AND CURRICULUM DESIGNER

    Callum is a recent Cambridge MMath graduate. He co-ran the first iteration of ARENA, and was head TA in Boston for 2023 Winter ML Bootcamp.

    He is a semi-regular climber, sci-fi movie fan, and the current world record holder for the most clothes pegs removed from a line and held in one hand (yes, really).

    This past year, he’s decided to pursue AI safety research fulltime (even though he probably peaked with the clothes pegs thing).

    You can read more of his writing here.

  • Kathryn O'Rourke

    OPERATIONS MANAGEMENT

    Kathryn is a recent graduate of King’s College London, where she was a university group organizer for Effective Altruism. She was operations director for the 2023 Winter ML Bootcamp in Boston.

Our TAs

  • Pranav Gade

    TA & CURRICULUM DESIGN: TRANSFORMERS, TRAINING AT SCALE

    Pranav is a student at IIIT Lucknow, studying CS&AI. He was a participant in MLAB 1.0, and a TA in Boston for 2023 Winter ML Bootcamp.

    He loves Minecraft, computers (especially if they are running Minecraft), reverse engineering, and reading sci-fi. You can find out more about him over at https://pranavg.me.

  • Arush Tagade

    TA & CURRICULUM DESIGN: TRAINING AT SCALE

    Arush is currently a grant-funded independent researcher working on vision model interpretability. He has previously been a TA for AGISF and Intro to ML Safety and was a SERI Summer Research Fellow in 2021.

    Arush has worked as a research engineer in the past at startups focusing on deploying computer vision models and speeding up model training through software optimisations.

  • David Quarrel

    TA: REINFORCEMENT LEARNING

    David is a PhD student from Australia on a research visit in Cambridge. He wrote some of the RL content for the first iteration of ARENA, and will be head TA during the RL weeks.

    He enjoys rock climbing and road cycling.

  • Joseph Bloom

    TA: FUNDAMENTALS, TRANSFORMERS

    Joseph was an ARENA round one participant and is now Independently Funded Alignment Research Engineer studying Mechanistic Interpretability of GridWorld Agents. He is also the current maintainer of TransformerLens, a popular open source package for mechanistic interpretability of transformers.

    Prior to working in AI Alignment, Joseph worked at as a data scientist at a computational biology startup.

  • Kay Kozaronek

    TA: FUNDAMENTALS

    Kay is a research engineer at NotodAI Research who’s on a mission to ensure advanced AI systems are as safe as a cushioned playground.

    With a background in management consulting and startup experience, Kay sees the creation of these safe AI systems as the most pivotal challenge of our time.

    His research focuses on the interpretability of large language models, a task as complex as figuring out what his partner wants for dinner.

    Kay values clear communication, honest feedback, and a good dad joke. If you ever need a pun to lighten the mood, he’s your guy. When he’s not deciphering AI or cracking jokes, you’ll likely find him immersed in sports, creating music on his piano or guitar, or simply spreading smiles.

  • Vinal Hiremath

    TA: TRANSFORMERS, TRAINING AT SCALE

    Vinay most recently completed a masters in computational neuroscience studying biologically-plausible machine learning models at ETH Zürich, and has since transitioned to working on AI alignment research as a more pressing problem that piques his interest. He last worked as a TA for the Machine Learning for Alignment Bootcamp (MLAB) in the summer of 2022.

    Otherwise, Vinay enjoys running and the occasional hike, along with spontaneous conversations about topics such as urbanism, public transport, and random geographical facts. He is also always glad to hear feedback about what may be working well or could be improved, in the course or otherwise.

  • Charbel-Raphael Segerie

    TA: CAPSTONE PROJECTS

    Charbel is co-Head of the AI Safety Unit at EffiSciences in Paris. He did research at the intersection of Cognitive Sciences and ML and previously served as the CTO of a small startup focused on building a search engine for academic papers utilizing the recent advancements in NLP. He led the development of the curriculum for the ML4Good bootcamps and served as a TA at Berkeley for the 2023 Winter ML Bootcamp.

    He loves Jazz Piano, Philosophy, hackathons, and when code looks like poetry. You can find out more about him over at https://crsegerie.github.io/.