Joost Sannen, Senior Software Architect at Sioux, explains: “Edge-AI brings enormous opportunities. Sioux helps realize them for our clients by combining our unique expertise in AI, mathware, software and electronics.”

INTELLIGENCE BEYOND THE CLOUD

Artificial Intelligence (AI) continues to evolve rapidly, enabling machines to take on complex tasks like learning, reasoning and decision-making. Traditionally, these capabilities rely on large datasets and compute-heavy algorithms running in cloud environments. But in many scenarios, waiting for the cloud is no longer viable.

"With Edge-AI, these trained models are used at the edge of the cloud (in mobile devices and local servers) to enable 'intelligent' behavior," Sannen explains. "We support our clients in embedding intelligence directly into their devices, enabling real-time behavior without relying on constant cloud access. When we talk about Edge-AI, we’re really talking about making embedded devices smarter."

WHY EDGE-AI?

"Do you want all your factory’s performance data uploaded to an American cloud server, or would you rather keep it to yourself?" asks Sannen. "Edge-AI minimizes data exposure and significantly reduces the attack surface for cyber threats, which is all the possible ways a malicious actor can access a digital system."

Latency is another key factor. Many AI applications must respond in milliseconds, think of defect detection in production, sound suppression in hearing aids or emergency braking in vehicles. "In such cases, it’s no help to first send the data to the cloud. Processing needs to happen in real time.”

Lastly, data centers already consume around 1% of global electricity, a number expected to quadruple in the coming years due to the rise of AI. "‘Edge-AI is inherently more energy-efficient than cloud computing, therefore more sustainable and cost-saving in operation," says Sannen. "It also reduces dependence on network stability, resulting in greater reliability." ”

THE BIGGEST CHALLENGES WE’VE SEEN SO FAR

Implementing Edge-AI for embedded comes with specific challenges. One major hurdle? Physical limitations. “Embedded devices often have limited computing power and may run on batteries,” Sannen explains. “So how can you still do something useful with AI? The technology isn’t designed for energy efficiency. Plus, clients often bring their own models, with architecture and training done in-house. That model and its data are core IP and can’t be shared. It’s an exciting challenge to bring such a model into production.”

Choosing the right hardware platform can also be a challenge. From Nvidia’s TensorCore to Hailo and NXP eIQ, every option has trade-offs in terms of speed, efficiency and compatibility. "It’s always a balancing act: how do we achieve high accuracy without driving up power usage or costs? That’s what makes our work so interesting."

HOW SIOUX BRINGS EDGE-AI INTO PRACTICE

At Sioux, Edge-AI becomes reality through Sioux AI Solutions: a dedicated group within Sioux Technologies. As a focused division, we develop practical, high-impact AI solutions that drive innovation for our clients. By combining deep RD expertise with hands-on experience, we deliver AI-driven breakthroughs that make a real difference.

Sioux AI Solutions focuses on the full lifecycle of AI development, from research and model training to optimization and deployment. We collaborate closely with Sioux’s software and electronics engineers when needed, especially when pushing the performance envelope with custom accelerators like FPGAs or tightly integrated embedded solutions.

A great example is Sioux’s work for Nemo Healthcare, whose NFMS system monitors fetal well-being using electrophysiological signals. This can help prevent delayed interventions when complications arise during pregnancy. With Edge-AI, Sioux managed to accelerate the client’s initial model by a factor of 100, enabling life-saving insights to be delivered immediately at the bedside.

LIMITLESS POTENTIAL

AI and machine learning aren’t new, but their practical use has exploded in recent years. "Almost every client we work with is either using or seriously considering AI," Sannen notes. "Edge-AI opens up real-time applications across sectors: production process monitoring, predictive maintenance, material analysis, agricultural monitoring and more. The possibilities are virtually limitless. Edge-AI is a revolutionary technology, opening up new opportunities across industries, and Sioux helps bring them to life."