Machine Learning Seminar Series Fall 2026 | | 3D Mechanical Shape as Code: Toward an Agentic Future for Engineering

Abstract: For decades, 3D mechanical design has relied on boundary representation (B-rep), a static description of geometry that underpins modern CAD, simulation, and manufacturing. While powerful, this paradigm has also produced engineering workflows dominated by graphical interfaces, manual operations, and disconnected software tools. This creates a fundamental bottleneck for modern agentic AI, as mechanical design and engineering still lack the programmable infrastructure needed for AI systems to interact with, reason about, and modify complex 3D designs. This raises a simple question: what if engineering design were treated as code? In this talk, I explore the idea of representing 3D CAD models as executable programs rather than direct static geometry. Programmatic representations expose the structure and logic of how designs are constructed, enabling automation and providing a natural foundation for training foundational generative AI models that can generate functional mechanical designs. I will briefly revisit earlier attempts at intelligent design systems and discuss how recent advances in artificial intelligence open the door to a new generation of AI-native engineering software. The discussion focuses on 3D mechanical design but points toward a broader future in which engineering workflows become programmable, generative, and AI-native.

Bio: Ferdous Alam is an Assistant Professor in the George W. Woodruff School of Mechanical Engineering at the Georgia Institute of Technology and an affiliated faculty member of ML@GT. He is also a core faculty member at the Institute for Robotics and Intelligent Machines (IRIM). He leads the Inference Lab, where his research focuses on the computational foundations of AI-driven engineering design and robotic manufacturing, including generative modeling of mechanical systems, representation learning for 3D geometry, and robot learning for manufacturing tasks. Before joining Georgia Tech, Dr. Alam was a postdoctoral researcher at MIT, and he previously received his PhD from The Ohio State University. His research has been recognized with several Best Paper Awards at leading ASME conferences, as well as the Google Research Scholar Award in Applied Science.

For CODA guest access, please contact shatcher8@gatech.edu at least 2 business days prior to the event.