NeoLens, an American deep-tech startup with Ukrainian roots, has unveiled the world's first AI assistant designed for the diagnosis and repair of military vehicles. This innovative system operates offline, making it suitable for use in combat zones. According to the startup's press release, the technology is already assisting mechanics in the Armed Forces of Ukraine to reduce the downtime of armored vehicles like the Humvee and MaxxPro.
Igor Simutin, a commander in the logistics repair unit, stated that the AI assistant saves mechanics valuable working hours, enhances their safety, and extends the lifespan of the equipment.
Currently, NeoLens is deployed in numerous repair battalions along the front line and is gradually being scaled to accommodate other weapon systems, including artillery and robotic platforms. The system is primarily utilized for Humvee repairs.
The startup was founded this year at Stanford University as part of the Lean Launchpad and Hacking for Defense programs by Ostap Korkuna, an American of Ukrainian descent, and James Leo, an American of Taiwanese descent. The developers emphasize that NeoLens is the first assistant to operate entirely without internet access, ensuring reliability and safety even in disconnected environments.
Ostap Korkuna noted that the solution emerged after hundreds of interviews with Ukrainian and American military personnel, who confirmed the need for an autonomous tool for vehicle repairs.
It is highlighted that NeoLens enables military personnel to perform step-by-step diagnostics, identify faults, and conduct field repairs at an expert level, even if the mechanic is using a specific platform for the first time.
The developers stress that the system is provided free of charge to the Ukrainian military.
The company plans to expand the functionality, including multimodal diagnostics via photos and videos, and to scale the technology for civilian sectors and the international defense solutions market.
Ukrainian military and volunteer repair initiatives can access NeoLens by applying.

6949 image for slide