UN Tech Over Hack-A-Thon
This Hack-A-Thon will offer a space for a collaborative innovation sprint where participants use open source tools and artificial intelligence to rapidly build solutions addressing global challenges aligned with UN goals.
Challenge 1: Safety, Supervision, and Governance in the Agentic World

Session led by:
Safety, Supervision and Governance in the Agentic World will be a challenge developed by AESIA (the Spanish Agency for the Supervision of Artificial Intelligence) and the Spanish Open Source AI Community, led by the Ministry for Digital and Civil Service Transformation of Spain. The aim of the challenge is to navigate the delicate balance required for responsible AI across its various dimensions as these considerations arise throughout development.
This challenge will call for expertise in the following areas:
• AI agent and software development (Python, open source agent frameworks)
• Fairness, privacy, and cybersecurity analysis in machine learning
• AI governance, human supervision, and auditability tools
Challenge 2: From Data to Action: Building Next Generation Evidence

Session led by:
How can we transform the world’s most valuable public data into tools that drive decisions for the Sustainable Development Goals and other pressing poly-crises?
This hackathon invites participants to build decision-ready data products using the UN system’s emerging Data Commons: a breakthrough knowledge graph platform with both MCP and SDMX API support that will host all the UN system’s public development-focused data and statistics in one place. Participants will create tools to turn diverse, real-world datasets into actionable evidence for humanitarian response, public policy, and development planning.
Working across an ecosystem of data, from the Data Commons to other publicly available sources from administrative data to frontier sources such as satellite imagery, climate risk models, and more, teams will design solutions that move beyond static dashboards toward dynamic, Agentic, decision-support systems.
This challenge will call for expertise in the following areas, depending on desired level of technical depth:
• Data visualization — Turning complex information into clear, accessible visuals and dashboards.
• AI‑powered applications — Building agentic systems that reason across interoperable data sources using APIs and emerging interfaces like MCPs.
• Multi‑agent orchestration — Coordinating multiple agents and data streams to produce real‑time insights.