Events

The eScience Institute is offering a 4-day workshop on introductory artificial intelligence (AI) and machine learning (ML) methods for structured data. The course will use […]

This quarter, in-person office hours for machine learning and AI will take place in person at the WRF Data Studio on the sixth floor of […]

This quarter, in-person office hours for data science will take place in person at the WRF Data Studio on the sixth floor of the Physics/Astronomy […]

This quarter, in-person office hours for machine learning and AI will take place in person at the WRF Data Studio on the sixth floor of […]

This quarter, in-person office hours for data science will take place in person at the WRF Data Studio on the sixth floor of the Physics/Astronomy […]

AWS offers access to a number of Foundation Models through its Bedrock service. This Cloud Clinic will demonstrate accessing the Claude LLM through the Bedrock […]

AI coding agents are reshaping software development practice at a pace that presents particular challenges for the research community. Existing resources are frequently fragmented, hype-driven, and rarely tailored to scientific research contexts. This workshop addresses that gap by providing researchers and research software engineers with a rigorous, practical introduction to AI coding agents and their application to research software engineering workflows. The session is organized into four blocks: an overview of the current AI coding agent ecosystem and tool selection considerations; a conceptual primer on how large language models are post-trained to support tool calling, planning, and iterative code generation; a live demonstration of an agent executing a complete research-plan-implement cycle aligned with Scientific Python community guidelines; and a guided hands-on exercise in which participants construct a simple agent themselves. The workshop prioritizes conceptual understanding over familiarity with any single platform, equipping attendees to evaluate emerging tools independently as the field evolves. Common failure modes in agent-assisted development are addressed explicitly, alongside practical mitigation strategies. No prior background in artificial intelligence or machine learning is required. Participants are expected to bring a laptop with terminal access and an active GitHub account.