Digitalization is a key step towards utilizing comprehensive process knowledge, which is required for enabling energy-efficient and sustainable manufacturing process chains. Injection molding of polymers is a highly scalable process; however, conventional manufacturing ramp-ups currently result in a significant amount of scrap before finding suitable process parameters. Machine learning-based data analysis is a versatile and indispensable tool for improving process parameters and enabling automated real-time quality monitoring. This research focuses on the development and use of thin-film sensors for in-process measurement of temperature and melt flow directly in the cavity. The sensors are directly integrated on tool surfaces, highly wear-resistant, and only a few micrometers thick. The study covers the integration of sensor inserts and fabrication using vacuum-based deposition techniques like physical and chemical vapor deposition. A small batch test showed that sensors combined with a machine learning algorithm on an edge device can determine real-time quality of each component.