Build an autonomous AI agent that can solve scientific research tasks — without knowing in advance what those tasks will be.
Each team designs a system — an agent scaffold, prompting strategy, tool-use pipeline, or any combination — capable of autonomously tackling data-driven scientific problems. You won't be solving the problems yourself; your agent will.
How it works
We provide a set of development problems: real scientific tasks with known solutions, drawn from experimental and observational data. Use them however you like — study the task format, run your agent against them, evaluate its outputs, and iterate on your design.
Final submissions are scored on a set of hidden test problems that no team has seen. Same format, different tasks. Your agent runs fully autonomously during evaluation — no human intervention allowed.
What we're testing
Can your agent generalise across scientific domains? Can it load unfamiliar data, form an approach, execute the analysis, and produce meaningful results entirely on its own?
Problem categories
Tasks are drawn from three domains: (1) Materials, (2) Science of AI / ML, (3) Climate, and (4) Bio. A fifth "super-category" — General Science — tests for cross-domain generalisation.
Program
1:30 PM - Doors open and check-in at Merantix AI Campus
1:45 PM - Hack kick-off and track intro
2:00 PM - 10 min pairing session
2:10 PM - Hacking begins
6:00 PM - Pizza & networking
7:00 PM - Winners announcement & showcase
8:00 PM - Networking
Location
Merantix AI Campus
Max-Urich-Straße 3, 13355 Berlin, Germany
Ask for directions at the reception
