During my internship at BMW, I addressed the rising demand for high-fidelity synthetic imagery used to train early object detection models. This project pushed me to adapt my industrial design background toward a digital-first approach, focusing on precise geometry for simulation rather than traditional manufacturing.

By applying hard-surface modeling techniques, I developed assets that accurately mirror real-world industrial components. This precision is essential for replicating "ground truth," requiring a strong grasp of topology and non-destructive workflows to ensure mesh integrity during rapid iterations. I also used a Physically Based Rendering (PBR) workflow to ensure light interacts with automotive materials like brushed aluminum and carbon fiber according to the laws of physics, providing the accurate reflectance values necessary for optical sensors.

Beyond individual assets, I utilized procedural tools to generate environmental variances—such as shifting weather and lighting—to help autonomous systems navigate complex "edge case" scenarios. Ultimately, this experience taught me to build scalable, synthetic data-ready assets and sharpened my technical agility in creating the simulation-ready environments that drive computer vision.