Cambridge AI4S
Join

A student-run community at Cambridge

AI for science, sorted by what works.

Machine learning now runs through engineering, chemistry, materials, physics and biology at Cambridge, but the expertise is spread across nine departments that rarely meet. We bring those people together to learn the shared toolbox, pair problems with methods, and sort every claim by the evidence behind it.

Join the community Read the ladder

Open to every department and every level. No machine-learning prerequisite.

What we do

Learn the toolbox once. Use it in your own field.

The same tools are being rebuilt in four buildings at once. We would rather learn them together, and spend the time saved on science.

  • Talks and journal club

    Fortnightly, rotating through the disciplines, with the evidence on the table.

  • Hands-on workshops

    Interatomic potentials, structure prediction, Bayesian optimisation and agents, on your own laptop.

  • Project matching

    Pairs of domain and machine-learning people on term-long mini-projects.

  • Directory and mentoring

    A maintained map of groups, courses and funding, and peer help with applications.

This term

Michaelmas 2026

Our first term's programme. Dates are provisional and go out on the mailing list first.

Michaelmas 2026 Full TermTue 6 Oct to Fri 4 Dec 2026

  1. Talk

    Founding meeting: AI for science, sorted by what works

    What has actually arrived, what has only been demonstrated once, and what is still a promise, followed by the routes in at Cambridge.

    Date to be announced

  2. Journal club

    Journal club: machine-learned interatomic potentials

    MACE-MP-0 and Matbench Discovery: what universal potentials get right out of the box, and where they still need fine-tuning.

    Date to be announced

  3. Workshop

    Workshop: protein structure prediction, hands-on

    Predict, inspect and question structures on your own laptop, and learn what a confidence score does and does not tell you.

    Date to be announced

  4. Talk

    Talk: learned weather forecasting and neural operators

    Why data-driven forecasting reached operations first, and what neural operators can and cannot do for PDEs.

    Date to be announced

  5. Social

    Project-matching evening

    Bring a problem or bring a method. We pair domain and machine-learning people for term-long mini-projects.

    Date to be announced

  6. Workshop

    Workshop: Bayesian optimisation in the lab

    Design experiments with a Gaussian process in the loop, on a real dataset.

    Date to be announced

  7. Talks

    Lightning talks from members

    Five minutes each on what you are building or reading. Propose one when you join.

    Date to be announced

Get in at Cambridge

The route in is specific: a module, a project, a group and a deadline.

Next deadline

  • University Postgraduate Funding Competition, first round

    Most PhDs, the MPhil MLMI, the BBSRC DTP and NanoDTC's first review. This round reaches the Cambridge Trust, Harding and departmental studentships.

All deadlines

Join

Bring a problem, or bring a method.

Open to anyone at Cambridge: undergraduates, graduate students, postdocs and staff, from any department. No machine-learning background needed.

Sign-up opens shortly

The mailing list and chat links will appear here as soon as they are live. Until then, everything on Get in at Cambridge is open to everyone.

How to take part

  1. Join the mailing list and the chat.
  2. Propose a talk or a dataset. Fifteen minutes is enough.
  3. Bring a problem to project matching, or bring a method.
  4. Volunteer as a track lead for your discipline.
  5. Tell us which groups and modules we missed.