Research and development in the field of machine learning has grown considerably within the past decade. As machine learning algorithms become increasingly unique and complex, there is a growing need for customizable, expressible, high-level representation of these algorithms. Furthermore, researchers and developers should have the ability to harness standard machine learning optimization techniques for their custom algorithms. We propose a demonstration of the Tile Embedded Domain Specific Language (EDSL), which provides users with a high-level programmable frontend to the PlaidML tensor compiler. We will show the Tile EDSLʼs ability to enable custom machine learning algorithm development while integrating the performance and portability benefits of a tensor compiler.