Tutorial on Agentive AI for Scientific Discovery
U1.105
Institut for Informatics, University of Bonn
The goal of this tutorial is to gain hands-on experience building AI agents for scientific discovery, with a special focus on the use of AI agents for drug discovery. Participants in the tutorial will learn how to build an original AI agent on a laptop (or PC) by connecting an LLM with drug-discovery software through MCP.
The tutorial takes place over two days:
- Day 1 (14:00-17:00, Thursday, 5 November): Participants learn how to develop foundational software tools for drug discovery using RDKit.
- Day 2 (14:00-17:00, Friday, 6 November): Participants learn how to connect the drug-discovery tools created on Day 1 to an LLM and use them through an MCP server.
Both days include hands-on exercises as well as challenges for participants. The tutorial is open to students and researchers at all levels.
The tutorial will take place in hybrid format. To participate in person at the University of Bonn, you need a computer science department login account (student or department member) for access to the department workstations, or else a computer or laptop of your own that is able to run a local LLM (<4GB) within LMStudio. Remote participation also requires access to a computer or laptop of your own with sufficient capabilities.
If you are interested in participating, please fill out the registration form at the link to the left.