Speakers
Description
Chairs: Peter Elmer, Sridhara Dasu
This tutorial provides a practical introduction to AI technologies that are transforming scientific research. Through hands-on exercises, participants will learn how to access and use large language models, work with multimodal data, build retrieval-augmented generation (RAG) systems over scientific documents, and develop AI agents capable of carrying out complex research tasks.
The tutorial will introduce modern agentic workflows for software development and scientific computing, demonstrating how specialized AI agents can collaborate to reason over information, generate code, and orchestrate multi-step tasks.
The session concludes with JFC (Just Furnish Context), a multi-agent framework for high energy physics analysis. Participants will see how an orchestrator and specialized subagents can collaborate across literature, software, data, and analysis tasks to support an end-to-end HEP workflow, illustrating the potential of agentic AI for scientific discovery.
All exercises will be carried out using resources provided by the National Research Platform (NRP), which offers managed access to advanced AI models through standard interfaces, allowing researchers to focus on applications rather than infrastructure.