In a groundbreaking initiative set to transform AI model training, Milestone Systems has announced Project Hafnia, a new platform designed to provide developers with access to high-quality, compliant video data. Leveraging NVIDIA Cosmos Curator and AI model fine-tuning microservices, the platform aims to democratise AI development by enabling widespread access to traceable and regulatory-compliant annotated video data.
Revolutionising AI Model Training
One of the biggest hurdles for AI developers worldwide has been acquiring accurate and reliable video data for training visual AI models. Project Hafnia addresses this challenge by aggregating compliant data from Milestone Systems’ extensive network of partners and customers. This data will be utilised to train AI models, enhancing their accuracy and efficiency in real-world applications.
Milestone’s initiative also integrates Vision Language Models (VLM), fine-tuned using NVIDIA’s AI microservices on a vast dataset of compliant transportation video. Optimized for NVIDIA GPUs and the company’s Video Search and Summarization (VSS) AI blueprint, this offering promises to streamline AI-powered video analytics and automation.
Unlocking AI’s Full Potential
“Artificial intelligence is our generation’s biggest game-changer. A major challenge for ongoing development is access to sufficient high-quality data,” said Thomas Jensen, CEO of Milestone Systems. “The Project Hafnia platform aspires to be the world’s smartest, fastest, and most responsible platform for video data and AI model training.”
The platform is expected to accelerate AI and analytics development significantly—up to 30 times faster than current standards. NVIDIA’s Deepu Talla, Vice President and General Manager of Embedded and Edge Computing, noted that “the next phase in development and adoption of visually perceptive Agentic AI services will be unlocked by recipes like NVIDIA VSS blueprint combined with widely available and accessible fine-tuned VLM models.”
Key Services and Applications
Project Hafnia will introduce two primary services:
- Training as a Service (TaaS): Developers can access high-quality, annotated video data to train AI models efficiently.
- Visual Language Model as a Service (VLMaaS): Targeted at smart city transportation and intelligent traffic services, this model is designed to improve operational efficiency across industries such as traffic management, manufacturing, airports, law enforcement, and business operations.
With improved annotation and expanded datasets, future AI-driven applications could reach unprecedented levels of accuracy, paving the way for large-scale automation in multiple sectors.
Transforming Traffic Management
One of Project Hafnia’s first use cases is a next-generation VLM optimised for traffic and transportation analysis. This model will support various applications, including traffic assessments, driving condition evaluations, alert validation, and incident reporting, helping cities manage infrastructure more effectively.
Currently, Project Hafnia is in its pilot phase, with early access available for developers at hafnia.milestonesys.com/joinwaitlist. Initially focusing on traffic-related video data, the platform plans to expand into other sectors over time, ensuring a broader impact in AI model training and deployment.
With Milestone Systems leading this charge, the future of AI-driven video analytics is poised for significant advancements, unlocking new possibilities in automation and intelligent decision-making.
