Scott Kemp

Beyond Predictive: IoT Architectures with SoftServe

How can smart maintenance, AI, and global IoT infrastructures be put into practice – despite labor shortages and complex machinery? Scott Kemp, Head of Manufacturing Services, EMEA, at SoftServe discusses these challenges with Ing. Madeleine Mickeleit, sharing insights from projects with Schunk, Continental, and NVIDIA. SoftServe demonstrates how scalable IoT backbones and AI applications deliver real value – for example, an AI assistant at Continental that reduces MTTR and boosts OEE by 10%. With Schunk, SoftServe co-developed an IoT backbone spanning the entire machine portfolio, enabling end customers to perform maintenance with the help of assistive functions. This comes to life in the practical example from OptoTech: Product Owner Vineeth Vellappatt offers a look into an AI-supported grinding process on the SM80 machine – including error detection, parameter analysis, and concrete recommendations for action. Technologically, SoftServe combines structured sensor data with unstructured knowledge (e.g. SOPs), embedded into a RAG model for fast information delivery – implemented across Microsoft Azure, NVIDIA Omniverse, AWS, and more. Standards like OPC UA and Unified Namespace lay the foundation for scalability. At the core: compensating for knowledge loss, empowering new workers, monetizing services – and turning AI from a buzzword into productive reality. SoftServe follows practical frameworks like “Double Diamond Thinking” and Proofs of Technology instead of just POCs. The episode kicks off with a short impulse from Onuora Ogbukagu (Deutsche Messe AG).