DIU contract advances rapid edge AI retraining for counter-drone systems. Vizgard and Syzygy Integration will demonstrate localized AI retraining capabilities designed to reduce response timelines from weeks to hours By Summer James / 11 Jun 2026 Follow UST The Defense Innovation Unit (DIU) has awarded a contract to Vizgard and Syzygy Integration LLC to demonstrate rapid counter-drone artificial intelligence retraining and optimisation directly on edge devices. Traditional methods for improving deployed AI models require extracting operational data, transferring it to analysts and data scientists, and waiting weeks for updated models. Under the new initiative, retraining and optimisation will be demonstrated on Syzygy's Snap-E-AI-RTR edge devices, ensuring no data leaves the device and removing cloud dependency. Operators will be able to flag performance issues from ATAK and iTAK within Syzygy's AwareX capability and deploy a retrained model within hours. The contract also includes sub-classification capabilities, enabling identification of the make and manufacturer of detected uncrewed aerial system platforms alongside payload alerting. This allows operators to move from basic drone detection to more detailed real-time intelligence. The effort builds on a partnership between the two companies that began in 2023, with Vizgard's FortifAI already deployed operationally across Syzygy's autonomous sensor programs. Wesley Mitchell, CEO of Syzygy, commented, "We've operated on the southern border and done counter-drone long enough to know that the model you deploy on day one isn't the model you need by month three. Being able to retrain and optimise directly at the edge on SNAP, without moving sensitive operational data off-device, changes how we respond as the threat picture evolves. Vizgard and Syzygy's integration makes that possible." According to the companies, deploying artificial intelligence without a retraining mechanism can leave operators relying on sub-optimal models that distract rather than support. By generating annotated and cleaned datasets directly within the operational system, the integrated workflow is intended to remove traditional processing bottlenecks and provide operators with tools to verify and retrain models at the edge. Edge-Based Visual AI Software Platform for Defence & Security Camera Automation