MVPro Vision Podcast: Integrating AI and Robotics

Eric Smith, CTO of Aether Advanced Technology, joined Davide Pascucci, CEO of Bright Intelligent Automation (Bright IA), as guests on Episode #010 of MVPro Media's Vision Podcast, hosted by Josh Eastburn. The conversation explored how the two companies are working together to bring AI-powered computer vision into real-world manufacturing and automation environments.

A Partnership Built on Complementary Expertise

Bright IA specializes in custom automation solutions for manufacturing, including PLC, HMI, and SCADA development alongside industrial networking. Aether Advanced Technology focuses on the AI side — building custom computer vision systems for industrial inspection, object tracking, and robotic guidance. The partnership combines Bright IA's robotics integration expertise with Aether's custom vision software to solve problems that off-the-shelf solutions can't address.

Topics Covered

The episode covered a wide range of subjects at the intersection of AI and industrial automation:

• Traditional vision vs. AI-powered approaches — when to use each, and how to avoid overengineering a solution
• Convolutional Neural Networks (CNNs) — how models learn to detect features and defects through trained filters
• Synthetic data generation — creating training data when real-world samples are scarce or expensive to collect
• High-mix, low-volume manufacturing — the challenges of dynamic environments where product variability is the norm
• Customer education — setting realistic expectations about AI capabilities and working collaboratively through implementation

Real-World Projects

Eric and Davide shared details on several recent projects, including welding tab inspection using robotic arms with integrated vision systems, and defect detection for small manufactured objects. They also previewed a live defect detection demo built for the upcoming Automation Stars of Texas Expo.

Key Takeaways

A recurring theme throughout the episode was the importance of breaking complex problems into manageable subtasks and choosing the right level of technology for each application. As Davide put it, the key is knowing when an application calls for a dynamic, adaptive AI approach versus a more traditional, deterministic solution.