Spatial AI Applications

3 units

3D computational methods are essential to modern design, engineering, geospatial, and medical applications. Both vibecoding and innovative new neural models are transforming the way those applications are conceived and developed. This project-based course will introduce key concepts, technologies, and stacks relevant to 3D and geospatial AI applications by enabling students to develop their own spatial apps. Course content will include image, point cloud, geometry, and mesh data structures, key spatial AI foundation models, custom model training and tuning, and development processes. Target platforms include mobile / web (three.js, kepler.gl), CAD (Rhino), and multiplatform (Unreal Engine). The class will culminate with presentations and demos of students’ original spatial AI applications. Familiarity with Python, Javascript, or C# is expected. Familiarity with Codex, Claude Code, of similar is helpful. Assignments will consist of three projects, two intermediate and a final project.