We will present a broad overview of UTRC’s Systems Department research in the area of machine learning to automate and optimize decision process for aerospace engineering, smart building and digital manufacturing. The presentation will show how we apply the state-of- art deep learning techniques in relevant industrial domains and then dive into our research in transfer learning with domain knowledge. At UTRC, we work with our businesses to build predictive models from historical data. Despite the hype around Big Data, obtaining more data in industrial settings is often expensive or time-consuming. Transfer learning (TL) reduces data requirements by leveraging data-rich source tasks to improve learning on different but related target tasks. Building domain knowledge (e.g., from engineers) into models can also reduce data requirements. We discuss different types of domain knowledge and how we can incorporate them into existing TL frameworks. We also touch on our domain adaptation work to address more general TL settings. We will conclude with research problems of interest to UTRC and discuss career and internship opportunities in the broad area of machine learning.