Kyiv, 30 July 2025 — During the AI Data Jam hackathon at Unit.City (Kyiv), the team of IT engineers and data analysts “Mine Watch AI” developed an AI model capable of detecting explosive hazards in drone imagery. As their model proved to be the most accurate and effective, the team received $2,000 in funding from UNDP in Ukraine and the Government of Luxembourg, and will continue work in collaboration with humanitarian demining experts.

The two-day hackathon brought together around 50 participants who processed approximately 30,000 images containing soil, metal, and vegetation, training AI to identify explosive hazards. These were real field images captured during the operations of HALO Trust Ukraine and UADamage.

The hackathon was organized by the Ministry of Economy, Environment and Agriculture of Ukraine, the Demine Ukraine initiative, the Humanitarian Demining Center, and the Ukrainian tech company UADamage, with support from the United Nations Development Programme (UNDP) in Ukraine and funding from the Government of Luxembourg.

Ihor Bezkaravainyi, Deputy Minister of Economy, Environment and Agriculture of Ukraine, noted the success of the event: “What began as an experiment has become a full-scale initiative with real results and international support,” Bezkaravainyi said. “Thank you to everyone who invested their time and energy into this vital effort. The motivation is immense: together, we’re building a unique product that can help demine Ukraine — metre by metre — without risking human lives, and with the power of technology.”

AI Data Jam builds on the earlier AI Labeling Sprint, during which participants contributed to creating an annotated dataset using aerial images collected during practical field tests conducted by UNDP, The HALO Trust, and UADamage. These images formed the training and testing foundation for the next stage: designing and improving AI-powered object detection models capable of identifying mines and other ERW amidst a challenging landscape of debris, scrap metal, and environmental “noise.”

Benjamin Lark, UNDP Mine Action Programme Manager in Ukraine, highlighted the importance of collaboration and accessibility in driving innovation in humanitarian demining: “Ukraine is demonstrating a remarkable level of innovation in humanitarian demining,” Lark noted. “What we are witnessing is the future being built in real time. Our goal is to make these innovations accessible to everyone working in this field. For UNDP, it is essential that this hackathon brings together all key stakeholders — government, veterans, tech teams, civil society, and international partners. This kind of collaboration drives practical solutions and strengthens the foundation for a safer and more inclusive Ukraine.”

The AI models and datasets generated through this series of hackathons will continue to be tested and refined, with the ultimate goal of creating an integrated software solution to support detection, analysis, and clearance processes in the field. Once operationalized, these tools could significantly accelerate mine action efforts, reduce risks for deminers, and serve as a scalable model for post-conflict recovery efforts around the world. 

Background:

The hackathon was organized by the Ministry of Economy, Environment and Agriculture of Ukraine, the Demine Ukraine initiative, the Humanitarian Demining Centre, and Ukrainian company UADamage, with support from the United Nations Development Programme (UNDP) in Ukraine and funding from the Government of Luxembourg.

Media Inquiries: Yuliia Samus, UNDP Ukraine Head of Communications and Advocacy, yuliia.samus@undp.org