10 June 2026 · Teaching Space 4 Room 301, One Pool Street, London, E20 2AF· 09:30–18:00
Eligibility: This hackathon is free to attend. Please note that at least one member of each team must be registered for the Fourth Workshop on Multimodal AI and present their hackathon project at the event on 11 June 2026. Click here to register for the workshop by 5 June 2026.
MultimodalAI'26 Hackathon registration are now open: Register here by 4 June 2026.
Full participant details — strand guides and test notebooks — will be sent to all registered attendees on 4 June 2026. Watch your inbox.
AI systems are being deployed in ICUs, autonomous vehicles, social housing, and robotic platforms — but can the people who depend on them actually trust them? Trust has to be earned with evidence, not assumed.
The MultimodalAI'26 Hackathon is a deployment-centric event. It invites registered workshop attendees to spend a day doing what most AI development skips: rigorously auditing whether AI systems that combine multiple data types are safe, fair, and honest about their own failures.
No single “best model” wins here. The strongest entry finds the most important real-world failure mode, explains it clearly enough for a practitioner to act on, and produces a structured report that could actually inform a deployment decision.
Most AI hackathons ask you to build the best model. This one asks you to prove it deserves to be trusted.
Working in teams of 3–4, you will spend a day evaluating AI systems across four consequential real-world scenarios — using synthetic datasets carefully designed to reflect genuine deployment conditions — and producing structured evidence of whether it should be trusted. You will not be judged on accuracy alone. You will be judged on evidence.
| Strand | Key question |
|---|---|
| 🏥 Clinical | Can you trust an AI in the ICU? |
| 🏠 Housing | Can you trust a dataset used for housing decisions? |
| 🤖 Robotics | Can your robotics evidence be trusted beyond the lab? |
Each strand provides guided notebooks, synthetic datasets, and parallel track roles so the work can be divided across your team. You do not need to be an AI researcher to participate — domain knowledge is as valuable as code.
Three teams will be selected as the winners of the following awards to receive prizes (Amazon vouchers) on Day 1 (11 June 2026) of the MultimodalAI'26 Workshop:
| Registration closes | 4 June 2026 |
| Date | 10 June 2026 |
| Time | 09:30–18:00 |
| Venue | Teaching Space 4, Room 301, One Pool Street, London E20 2AF |
| Eligibility | Each team must have at least one member who is a registered attendee of the MultimodalAI'26 workshop and present at the event on 11th June 2026. |
| Team size | 3–4 participants. Solo registration is welcome, with team formation on the day. |
| Co-located with | Fourth Workshop on Multimodal AI, 11–12 June 2026 |
Full strand details and test notebook with detailed instructions will be sent directly to registered MultimodalAI'26 workshop & hackathon attendees on 4 June 2026.
If you have a question in the meantime, contact the organising team: multimodalai26-group@sheffield.ac.uk
The hackathon is open to everyone, including graduate students, postdoctoral researchers, academic and research staff, industry researchers and engineers, and domain specialists such as clinicians, housing officers, and safety inspectors. No prior machine learning experience is required, especially if you are working in a team with technical members.
At least one member of each team must be registered for the MultimodalAI’26 workshop on 11 June. Other team members may register for the hackathon only and are not required to attend the workshop.
Registration closes on 4 June 2026. Strand guides and starter kits will be sent to all registered participants on the same date.
Demo presentations, winner announcements, and prize awards will take place on 11 June 2026, during Workshop Day 1.
Yes. Webinar attendees and registered participants will receive the webinar slides and other useful links.
No. The hackathon is designed for cross-disciplinary teams. Domain expertise is as valuable as technical expertise. If you understand the clinical, housing, robotics, or deployment context, your contribution will be important. Winning teams are likely to combine technical analysis with clear, plain-English reasoning.
No. Solo participants are welcome. Team formation will take place from 10:20 to 10:30 on the day. You may also arrange a team in advance through the UKOMAIN Discord channel. To join the UKOMAIN Discord channel, please click here.
You can introduce yourself in the UKOMAIN Discord channel. Share your background, preferred strand, and any relevant technical or domain expertise. This is encouraged but not required. Formal team formation will still take place on the day.
No. It is optional. Introducing yourself on Discord may help you connect with others before the event, but team formation time is built into the programme.
Before 10 June, please:
That is fine. You may use any code editor or IDE you are comfortable with, including VS Code, PyCharm, Jupyter Notebook, or Google Colab. If you are not coding, you can still contribute through interpretation, reporting, domain reasoning, usability testing, or presentation.
Please bring:
Eduroam or guest Wi-Fi details will be provided on the day. Lunch and refreshments are included.
No. All strands will provide synthetic or prepared datasets in the starter kit.
Participants should use the datasets and starter materials provided for their chosen strand. Where appropriate, teams may use their own tools, methods, or models, but these must be clearly documented in the submission.
Yes. GitHub Copilot, Claude, ChatGPT, and similar tools are permitted. You must disclose any AI tools used in your submission. Judges may ask questions during the demo to confirm that your team understands what was built and why.
No. The challenge strands and suggested directions will be included in the strand guide. Your team will review the options at the start and choose a direction together. Domain specialists do not need to arrive with a prepared technical problem. Their role in interpreting findings and making recommendations is highly valued.
Team formation runs from 10:20 to 10:30. If you arrive after this, speak to an organiser immediately. They will help match you with a team that has space or with other late arrivals.
Building begins at 11:00, so a short delay is manageable, but the sooner you join a team, the more you can contribute.
No. Teams are formed during the dedicated team formation slot. Organisers will support solo participants in finding suitable teams.
Teams should ideally have two to four people. A strong team may include:
No. Each participant must join one team and work on one strand for the full day.
Your role is highly valuable. You can help the team interpret results, assess whether findings are realistic, write plain-language recommendations, identify practical risks, and make the final deployment verdict credible to real-world users. The judging criteria reward clarity, deployability, and the ability to communicate findings to non-technical decision-makers.
Probably not. This hackathon focuses on evaluating, auditing, and explaining AI systems, rather than training large models from scratch. While you may choose to train lightweight models, most tasks should run comfortably on a standard laptop. If your team does need GPU compute for model training, you’re welcome to use Google Colab or any other cloud-based resource you prefer.
Yes. Google Colab may be used, especially if your team needs additional compute.
There is no required IDE. You may use any tool you prefer, such as VS Code, PyCharm, Jupyter Notebook, Google Colab, or another editor.
Submission is via GitHub. Your team should push its work to a team branch and open a Pull Request against the relevant strand branch before the deadline.
Use your team name in lowercase, with hyphens instead of spaces. Examples: team-alpha, cardiff-med, ucl-robotics-2. Do not use special characters.