Fast-Tracking Edge AI Development: Insights from the EmbedUR Webinar

Edge AI is quickly becoming one of the most exciting areas in tech, but for many teams, turning an idea into a working product is still a big challenge. That was the focus of a recent Wevolver webinar titled "Edge AI Development: From Concept to Deployment," hosted by Samir Jabber and featuring Eric Smiley and John Marcconi from EmbedUR Systems. 

The session offered a deep dive into ModelNova, a platform designed to speed up and simplify edge AI development. But more importantly, it unpacked some of the key challenges and exciting shifts happening in the world of intelligent edge devices.

What Is Edge AI and Why Does It Matter?

Edge AI refers to deploying AI models on local devices, like cameras, microcontrollers, or sensors, rather than relying solely on cloud computing. This approach enhances privacy, reduces latency, and decreases network load. 

As Eric Smiley noted, 

“It’s about making everyday devices smarter and more responsive, without compromising privacy.”

According to John, another challenge is fragmentation. With so many hardware platforms, tools, and model formats, just getting a prototype running on a low-power device can take weeks or even months. That’s where ModelNova offers a solution.

What is ModelNova?

ModelNova was built to simplify this process. It is a full-stack solution that helps developers, from hobbyists to large OEMs, bring AI applications to life on edge devices. According to Smiley, the idea came from a noticeable gap in the ecosystem. While many open-source AI models were available, most were designed for powerful GPUs, not for small, embedded devices. 

To solve this, EmbedUR started converting and optimizing these models to run efficiently on edge hardware. The result is a platform that offers pre-trained, pre-optimized models for common use cases such as computer vision and audio detection. With Model Nova, developers can go from idea to prototype in minutes instead of weeks.

What makes ModelNova stand out?

• Pre-ported Models: Choose from 70+ optimized models tailored for small devices. No need to convert formats or spend weeks trying to fit a model into limited memory.

• ‘Try It Now’ Feature: Test computer vision models directly in your browser before moving to hardware.

• Detailed Benchmarks: See real performance and power usage data across supported devices.

• Blueprints: Step-by-step guides for common use cases like face recognition or trash sorting.

This is all aimed at reducing the time it takes to go from idea to proof-of-concept from months to minutes.

Real-World Use Cases

Many open-source models are trained on large, general-purpose datasets like Microsoft COCO, but real-world needs are often much more specific. Think: identifying defective screws on a factory floor or monitoring wildlife in a backyard.

Eric and John addressed this by highlighting Model Nova’s support for custom and synthetic datasets. Teams can start with public models, retrain them on unique data, and even use generative AI tools to expand small datasets.

As Maronei explained, You may not find a public dataset for your unique manufacturing screw, but you can create one, or even generate synthetic data to train your model.

Edge AI Is Not Slowing Down

Looking ahead, Eric and John see the Edge AI space growing beyond vision and sound, into real-time interaction, personalization, and privacy-preserving intelligence, all running on-device, without needing to send data to the cloud.

New use cases are imagined daily: smart coffee makers that know your preferences, home security systems that react locally, and even AI narrators that turn your backyard wildlife into a mini nature documentary.

And it’s not just companies driving this. As John pointed out, the real innovation may come from students, hackers, and small startups who now have access to tools that used to be locked behind enterprise doors.

Want to Try It Yourself?

ModelNova is free to explore. All you need is a Raspberry Pi to get started. Just head to modelnova.ai and sign up for an account. Whether you’re a pro developer or just curious, there’s something for everyone, and maybe a few surprises too.

Final Thoughts

This webinar was a powerful reminder that the Edge AI revolution is already underway, but it is only just getting started. If you want to participate, this is the time to dive in. Watch the full webinar to get all the insights, stories, and technical details that didn’t make it into this summary.

Watch the full webinar here.