Manufacturing generates massive amounts of data, yet many factories still run expensive machinery on settings that have not changed in a decade. Daniel Kazani sits down with Dr. Jonathan Spitz, Founder and CEO of GaussML, to discuss why having data does not always mean having information.

ㅤ

Jonathan explains his "small data" approach to industrial optimization. Instead of requiring months of data cleaning and massive data lakes, his team focuses on rapid experimentation. By running a few targeted tests, operators can find the ideal parameters for processes like laser cutting and injection molding in a single day. Jonathan shares real-world examples, including how a 0.5-gram adjustment saved Coca-Cola 20 tons of plastic a year and how job shops eliminated Saturday shifts by increasing efficiency. The conversation also covers the role of the human operator as a pilot rather than a bystander.

ㅤ

Guest Bio

Dr. Jonathan Spitz is the Founder and CEO of GaussML. Before launching his own company, he served as a Research Scientist at the Bosch Center for Artificial Intelligence, where he applied machine learning algorithms to industrial optimization. He holds a PhD in Mechatronics, Robotics, and Automation Engineering from the Technion - Israel Institute of Technology. Jonathan specializes in "small data" solutions that help manufacturers improve efficiency without complex integration.

ㅤ

What We Cover

1. The difference between being data-rich and information-poor in manufacturing
2. Why traditional deep learning often fails in factory settings due to the need for massive datasets
3. How the "small data" approach works: finding optimal machine settings with minimal experiments
4. Real-world wins: Reducing cycle times by 50% in machining and saving raw materials in bottle production
5. The Coca-Cola case study: How a tiny weight reduction per bottle resulted in massive material savings
6. The "Copilot" philosophy: Why AI should augment the operator's intuition rather than replace it
7. Overcoming the "worker gap" by making expert-level machine operation accessible to newer employees
8. Why is failing during the testing phase necessary to find the true limits of a machine

ㅤ

Resources Mentioned

1. GaussML (Official Website)
2. Optimyzer (Product)
3. Dr. Jonathan Spitz (LinkedIn)
4. Daniel Kazani (LinkedIn)
5. Softup Technologies
6. Bosch (Company)
7. TRUMPF (Company)
8. Coca-Cola (Company)