The Department of Defense (DoD) is seeking proposals for the topic "TAK Mobile Machine Learning (MML) Model Development" as part of the SBIR 23.3 BAA. The objective of this topic is to develop and train cutting-edge machine learning models for edge deployment via TAK (Tactical Assault Kit) using the Model Integration Software Toolkit (MISTK) format. The technology is restricted under ITAR or EAR regulations. The training can be done server-side, but inference must be done on the device. Offerors are provided with TAK-ML, a client and server-side framework for ML development, and NodeDrop, a technology to reduce the size of neural networks. Sample models/algorithms developed in and integrated with TAK-ML are provided. The potential applications of this technology include geolocation, command and control, search and rescue, surveillance, communications, IoT, cloud, and intelligence. The project will have a Phase II, which includes data collection, model design, implementation, training, testing, and evaluation at the tactical edge. Successful Phase II technology development may be eligible for additional Phase III work, with specific transition paths depending on the domain and problem set selected. The proposer will work with the Tactical Assault Kit (TAK) Product Center and end-user communities to promote the transition of machine learning models. The project is open for proposals until October 18, 2023. For more information, visit the DoD SBIR 23.3 BAA on grants.gov or the DoD SBIR/STTR website.