From Fall 2009.
I love the first project in this course - graph-based searches and performance testing over unbounded domains. A* versus Best-First was a particularly interesting subtle battle. I was still new to Java. I look at it and would love to give it a quick rewrite. The constraints of no CLA's or alteration of some boilerplate code made for a few awkward switches. I believe I would handle such constraints more gracefully now.
The class complained that the projects took too long, and the course was assigned 4 credits thereafter.
This class was foundational for my Master's degree - I was hired as the professor's research assistant, and so continued to be be inspired by his great understanding of AI, and completed the "Plan A" research track of the MS program under his direction.