Matthias Loeschl

AI Summary

→ WHAT IT COVERS Matthias Loeschl from Siemens Factory Automation explains how the Siemens-NVIDIA partnership brings AI solutions to manufacturing floors, addressing the 92% AI skills gap and 70-80% project failure rates through democratized tools and standardized infrastructure. → KEY INSIGHTS - **InSpecto Quality Inspection:** Out-of-box AI system trains on 20 good product samples in under one hour, requires no computer vision expertise, and enables operators to independently configure visual defect detection across metal, plastic, and electronics manufacturing without data scientists. - **Scaling Challenge Solution:** Siemens Industrial AI Suite closes the complete MLOps cycle from cloud training to shop floor deployment, achieving 25-fold acceleration in AI execution for Audi's automated weld spot inspection across 5 million daily welds using NVIDIA L4 GPUs in industrial PCs. - **On-Premise Copilots:** Industrial copilots run locally on shop floor using NVIDIA NIM microservices to address security concerns, enabling operators to query real-time production data and maintenance documents without sending sensitive factory data to cloud environments for faster decision-making. - **AI Adoption Barriers:** Studies show 40% of manufacturers distrust AI, 92% lack skilled experts, and only 16% achieve AI goals because companies build from scratch instead of using standardized infrastructure, causing most projects to fail at proof-of-concept stage without production rollout. → NOTABLE MOMENT Siemens demonstrates robots using foundational models to grasp arbitrary unseen objects for warehouse picking, eliminating the need to pre-program handling for each item shape, though simulation-to-reality gaps still require real-world data for fine-tuning before deployment. 💼 SPONSORS None detected 🏷️ Industrial AI, Manufacturing Automation, Computer Vision, MLOps