Delaware's state government held its first AI hackathon this week, pairing more than 40 university and high school students with 10 state agencies to build machine learning tools for real government problems. The three-day event, called "Hennovate the State," ran at the University of Delaware's STAR Campus and focused on challenges like records management, customer service, transportation planning, workforce benefits and internal reporting.
Students from the University of Delaware, Penn State and area high schools worked alongside state employees on case studies submitted by agencies. The event was the fifth hackathon hosted at the university, but the first to include state agencies and government datasets, according to the Delaware Department of Technology and Information.
Agencies brought real problems
Delaware Chief Information Officer Bob Osmond, who took the role in April after serving as Virginia's CIO, said the format let agencies tap into skills they don't always have in-house. "Our agency partners brought real problems to the table, and these students brought fresh thinking and sharp technical skills," he said in a press release. "That's how we find AI solutions that actually work for the people of Delaware."
Teams received mentoring from Amazon Web Services, Google, Microsoft and Nvidia throughout the event. At the end, judges recognized four projects for impact, technical depth, creativity and problem execution. The most promising concepts may be evaluated for possible development, DTI said.
The hackathon builds on Delaware's existing AI work. Last year, the state launched the AI Sandbox Program to experiment with machine learning for traffic monitoring and network intrusion detection. Delaware joins California, Georgia, Texas, New York and Utah in establishing sandboxes or testing environments for AI, which allow agencies to pilot tools, study governance practices and get temporary relief from specific regulations. For government employees, these programs offer a path to test AI for AI for Government applications without committing to full deployment.
What students produced
The four winning projects were selected from teams that spent the three days refining their tools with agency feedback. DTI did not release specific details about the winning solutions or which agencies submitted the case studies.
The event's structure - short timeline, real datasets, agency mentors - reflects a growing approach among state governments to use hackathons as a low-cost way to explore AI. Rather than issuing a request for proposals or hiring consultants, agencies get working prototypes and students get experience with public-sector problems.
Why this matters for government professionals
For state and local employees, Delaware's approach offers a template: bring a specific problem, pair it with outside technical help, and see what comes back in days, not months. The sandbox program and hackathon both point to a broader shift - agencies are testing AI in controlled settings before making procurement decisions. Government professionals who want to participate in similar efforts can start by identifying a narrow, well-defined process problem in their own agency and asking whether a pilot or sandbox program already exists to address it. For those in policy roles, understanding how these testing environments work is becoming part of the job, since regulatory sandboxes are increasingly the mechanism states use to shape AI governance. Training resources like AI for Policy Makers can help officials evaluate these programs with a clearer sense of what they can and can't deliver.
Your membership also unlocks:
