Manufacturing companies have embraced Internet of Things and data analytical technologies to optimize manufacturing processes. In a typical case, they benefit from having a controlled environment with well-structured workflows and established facility layouts. However, production of highly customized products and organization of shop-floor operations remains a challenge. This project investigates the problem of work-in-progress inventory management in the case of on-demand production of customized products having a multitude of production workflow variants and a significant share of human operations. It aims to create an information system, which uses Internet of Thing to track movement of work-in-progress inventory and optimizes materials picking and placement. The computer vision is used to reduce the need for materials tagging and spatial database technologies are used to create a dynamic view of the shop-floor layout taking into account current locations of work-in-progress inventory. The system can be customized for various manufacturing facilities. The project is carried out in a university industry collaboration involving companies specializing in industrial Internet of Things. The system is tested at a medium-sized printing company