CSCI 7000-09 Fall 2026: Building AI Agents

Meets Remotely on Zoom: https://cuboulder.zoom.us/j/99193569574 (you need to authenticate with your colorado.edu credentials)

Three times per semester (Sept 9th in ECCS 1B14, mid semester, and last week of the semester), this class meets in person on campus.

Prof. Danny Dig (danny.dig@colorado.edu)
Office: On Zoom
Phone: (303) 492-1293
Office Hours: right after class, additional slots offered each week to accomodate various schedules.

Students should have basic knowledge of Computer Science undergrad classes, including software engineering, programming languages, systems, ML/AI as demonstrated by undergrad courses at CU. This is NOT an introductory grad course to Software Engineering, for those students we recommend other grad courses such as Foundations of Software Engineering (CSCI 5828). This course assumes you are a competent software engineer already -- we are looking into how you can augment your current skills to include GenAI supported software development. If you are not sure whether you can attend this course, please consult the instructor (email: danny.dig@colorado.edu)

The purpose of this research course is to expose students to seminal topics and recent trends in the field of GenAI-powered Software Agents.

This graduate-level course explores the rapidly emerging paradigm of building software systems from agentic components -— autonomous, reasoning, and goal-directed software entities that can perceive context, collaborate, and take initiative. The course has two central goals: (i) to critically read and discuss state-of-the-art research papers on software agents, multi-agent collaboration, and agentic AI in software engineering, and (ii) to gain practical experience designing, implementing, and evaluating an agentic system.

In this highly interactive, seminar-style class, students will engage in instructor-guided discussions, student-led paper presentations, and hands-on labs exploring current agent frameworks and architectures. The course will culminate in a team-based term project—either research-oriented or prototype-driven—that demonstrates creativity, technical rigor, and a deep understanding of how agentic components can transform modern software systems.

Our exploration consists of in-class discussions of the cutting-edge research literature, interactions with industry thought leaders, and team projects that encourage risk taking.

The course offers an opportunity for students to pursue research-oriented or novel industrial term projects in small teams (3-4 students). Students may pursue any project of their choosing that explores building software systems powered by Agentic components.

The course offers several opportunities to learn from the movers and shakers in Agentic AI from industry and academic research labs.

COURSE OBJECTIVES

• expand student's knowledge more broadly about Software Agents research & practice through reading research papers and industry surveys
• develop critical thinking abilities by being able to assess the quality of published research and pose new research questions
• practice giving scientific presentations and teaching others
• engage in active learning activities in class, such as student-guided discussions on research and applications of agentic AI
• practice developing a research or novel-industrial software development project through all its stages: formulating a research problem, posing research questions, gathering related work, designing a solution, evaluating the solution empirically, writing a research paper
• practice engaging in dialogue with thought leaders in agentic AI.
• have fun learning