We caught up with Francisco Melo to know more about the spin-off Binedge.ai, a Portuguese company that intends to deliver EdgeAI solutions for companies, users, and developers.
This brand-new spin-off brings a revolutionary answer based on EdgeAI. It creates and displays computationally expensive AI resource-constrained for small devices. EdgeAI solution enables companies and developers to quickly develop high-performance and ultra-low-power deep-learning algorithms in the device itself, eliminating the need for constant connectivity infrastructure (no need to send data continuously over to a central computing server) while safeguarding data privacy.
This innovative technology will disrupt the EdgeAI industry by offering 11x more performance and 75x more energy efficiency than existing solutions in the market.
- Who's behind this idea?
Everything started with my master's thesis, under the supervision of Hugo Silva (IT-IST) and Horácio Neto (INESC-IST). Now, the founding members are Francisco Melo (me), Hugo Silva and Rui Maciel.
- What is the research behind this new spin-off?
Research was developed in a new area of artificial intelligence called EdgeAI, whose objective is to create high-performance solutions to be able to run AI algorithms on very small devices. Therefore, AI is something, in computational terms, a little complicated and difficult to implement outside the classic paradigm of large computers, with great computational power and energy consumption.
- What services will be provided?
What we do is provide solutions, which allow users to run their AI-based applications on microdevices with low computational power and low energy consumption.
- Was this spin-off the result of your Ph.D. or Masters's research?
The spin-off resulted from my master's research. I worked here at the Instituto de Telecomunicações from my fourth year until about a month ago. Developing research in a slightly different area, with Prof. Hugo Silva, more in the acquisition of biosignals, and in parallel he was also there working on this technology. So it was complimentary.
- What does EdgeAI technology mean?
It is not a technology, but more an area of research and industry, which specializes in applying AI to microdevices.
- And at this stage, what next?
Finally, I was able to dedicate myself time to the start-up, of this project, having already the company already formally constituted, and now we are on the way to launching the first product, which is an online platform that allows users to easily introduce their algorithms of AI on the small devices, but also create and develop your AI models. It is a win-to-win solution that accompanies you from design to AI development, to its implementation on small devices.
- Is that an app?
No. We develop the tools for users to be able to easily run these apps with very efficient AI.
- Is this spin-off goal to end up supporting and improving what already exists?
Exactly. We support and optimize AI for small devices, namely mobile phones, and more customized embedded systems, for example, hardware that a company is making for a specific application. It could be a chamber in an industrial environment.
- What was the solution like before?
For now, AI is very recent. Like Edge AI.
Before, we were always dependent on large machines, due to processing needs and energy consumption. But now if a person or company needs to develop a product closer to the data source, for example, a camera that can identify a person, a question already arises: how can we process AI in such a constrained environment in terms of resources? And, we have the answer.
- So, can you say that you offer a consultancy service?
Yes, it is a kind of an automated consulting.
- In addition to this route you have taken, do you intend to go back to academic research?
I never say never, but right now this project is demanding a lot from me, which I need to focus 100% on. In the future, who knows?
Visit the start-up webpage: