Skip to Content

Milestone CEO emphasizes domain-specific models for success

Milestone CEO emphasizes domain-specific models for success ‘In the future, we could expand into law enforcement and public safety,’ Thomas Jensen says

Milestone CEO emphasizes domain-specific models for success

COPENHAGEN, Denmark — Milestone Systems CEO Thomas Jensen believes Project Hafnia will have an immediate impact on traffic and transportation datasets across diverse city environments, but he says the longer-term impact of the technology is much broader. 

“Rather than pursuing general AI, we believe in an agentic approach—smaller, domain-specific models that excel in their respective fields,” he said. “Our long-term vision is to dominate the traffic and transportation AI domain while enabling partners to leverage Hafnia’s capabilities in other sectors.” 

In part one of this interview, Jensen focused on the importance of ethically sourced data. In this second part, he explores the future of Project Hafnia and how he sees the technology branching out. 

SSN: How do you envision Project Hafnia transforming industries such as security, transportation, and urban planning? 

Thomas Jensen: That’s a broad question but let me break it down. 

For transportation, Project Hafnia democratizes traffic pattern optimization for cities worldwide. Cities and their technology partners can train AI models on their specific urban data to optimize traffic flow. If a city lacks the resources or video data, we provide fine-tuned models based on comparable datasets, ensuring ethical and anonymized data usage. This can help reduce rush-hour congestion, optimize real-time traffic management and even lower pollution by minimizing idle vehicle emissions. 

For urban planning, Hafnia can support the development of new neighborhoods with optimized traffic patterns, efficient lighting and smart city infrastructure. By analyzing data from multiple cities, planners can make informed decisions about urban development and integrate real-time traffic management solutions. 

Regarding security, Hafnia initially focuses on traffic and transportation models. However, in the future, we could expand into law enforcement and public safety, helping to reduce crime rates by developing models trained on ethically sourced security data. While traffic and transportation remain our current priorities, broader applications are on the horizon. 

SSN: Could you discuss the role of NVIDIA’s NeMo Curator in enhancing Project Hafnia’s data curation and AI model training capabilities? 

Thomas Jensen: Absolutely. I’ll keep it simple since I’m not a computer vision developer myself. 

NeMo Curator significantly accelerates our ability to deliver highly curated, large-volume data. Its software stack is designed for scalability, making our curation process more efficient. Additionally, it enables co-creation of a customized pipeline tailored to our specific data needs. In essence, NeMo allows us to focus on our strengths without needing to develop the entire software stack from scratch. 

SSN: Looking ahead, what are the next steps for Project Hafnia? How do you plan to expand its services and data offerings? 

Thomas Jensen: Our immediate priority is expanding traffic and transportation datasets to enhance model accuracy across diverse city environments. Currently, Hafnia is invite-only for two reasons: First, we want to scale responsibly and avoid the pitfalls of overextending too quickly, as many AI companies have done; second, we aim to refine and solidify our initial traffic-focused models before expanding further. 

The next step involves allowing technology partners to develop applications trained on Project Hafnia, making these apps accessible to customers worldwide. Eventually, we plan to broaden our reach beyond traffic and transportation to industries like law enforcement, hospitality, retail, and health care. The goal is to create specialized AI models that deliver high accuracy within their domains. 

SSN: Great insights. Is there anything else you’d like my readers to know? 

Thomas Jensen: I’d like to emphasize the importance of ethically sourced data and the right of every citizen to control their own information. Our mission is to ensure AI development respects privacy, human rights and democratic freedoms. That’s at the core of everything we do. 

Comments

To comment on this post, please log in to your account or set up an account now.