AI developers across 15 countries collaborated with Milestone Systems to shape the future of safer, more efficient video intelligence at the 2025 Milestone Developer Summit, where 55 innovators received early access to Milestone’s upcoming Vision Language Model (VLM) powered by NVIDIA.
A Global Summit for Innovation
Hosted at the historic Carlsberg Museum in Copenhagen—a site known for groundbreaking scientific discoveries—the summit brought together developers, technology partners, and industry leaders. Keynotes from NVIDIA, AWS, Dell, and Intel explored future trends in open platform innovation and responsible AI.
Sebastian Döllner, VP of Technology Partnerships & Open Platform, emphasized the importance of collaboration:
“Technology is evolving faster than ever, and no single company has all the answers. The essence of Milestone’s open platform is to empower innovation.”
Early Access to Milestone’s Specialized VLM
Before the summit, selected AI developers tested Milestone’s upcoming specialized VLM. Built using NVIDIA Cosmos-Reason and refined with Hafnia’s domain-specific data library, the VLM is designed to interpret complex city environments, languages, symbols, and weather conditions.
The Hackathon challenge tasked developers to create third-party integrations and leverage the VLM API to improve smart-city operations.
Driving Smarter, Safer Cities
Milestone also previewed a new generative AI plugin for its XProtect Video Management Software. This tool automatically transforms video content into summaries, written reports, and validated real-time alerts—supporting applications in traffic management across cities, ports, airports, and large urban spaces.
Hackathon Winners Announced
Six finalists presented their solutions, and a panel from Milestone and NVIDIA selected the top three.
Winner: Thomas Kreutz — Ask The City
Using Hafnia’s VLM API, Kreutz’s solution converts real-time city camera footage into immediate, privacy-aware answers to natural-language questions.
“I wanted to build something relevant and useful in real life,” he said. Kreutz received €5,000 and the NVIDIA Jetson AGX Thor Developer Kit.
Roland Harwood, Community Lead – Hafnia, noted:
“Thomas Kreutz and Ask The City really nailed it. It’s a promising demonstration of how our platform and curated data library can support the training of secure and anonymized computer vision models.”
Finalist Ranking
- Ask The City — Thomas Kreutz
Real-time answers generated from live city cameras using Hafnia’s VLM API. - VLM for Emergency Response — Flynn Owen, Dan Towers & Tom Aldridge (Blaize)
Edge-AI solution generating events and response plans in near real-time. - SmartMap — Gabriel Crosnier de Bellaistre (Citilog)
Unified traffic, visibility, and incident information visualized on live maps. - RevoFlow — Rawinder Singh
A no-code, drag-and-drop workflow builder for AI video analytics. Audience Award Winner. - Clipnotes — Magnus Guldberg Pedersen
Converts video segments into structured analytical insights. - Hafnia Video Event Analyzer — Jojy Saju JosephS
Multimedia intelligence platform enabling semantic metadata creation and event detection.
Milestone Systems is a global leader in data-driven video technology, delivering solutions for industries including airports, manufacturing, traffic management, retail, and law enforcement. Its portfolio—spanning XProtect VMS, BriefCam analytics, and Arcules cloud VSaaS—helps organizations understand the past, monitor the present, and anticipate future events.
Generative AI for Video Analytics.
Frequently Asked Questions (FAQs)
What is Milestone’s new Vision Language Model (VLM)?
It is a specialized VLM trained on ethically sourced video data and enhanced with Hafnia’s domain-specific city datasets.
Who participated in the Hackathon?
AI developers from 15 countries, selected for early technical access.
What was the winning solution?
Ask The City by Thomas Kreutz, which turns city cameras into real-time, privacy-aware intelligence.
How will the generative AI plugin for XProtect help cities?
It will automate video review, generate summaries, and deliver validated real-time alerts to improve traffic and urban management.
Where was the summit held?
At the Carlsberg Museum in Copenhagen, a venue known for historic scientific achievements.