How Medios Technologies’ Florian Savoy prepares mentees for the ‘messy’ world of datasets and AI

I’d spent over five years undertaking research in computer vision, a field of artificial intelligence (AI) that trains computers to understand and interpret images and videos. But I wanted to transition away from academia and do something more impactful and meaningful. 

With the encouragement of a local investor programme that supported technology entrepreneurs, I co-founded Medios Technologies in 2018. Together with fellow entrepreneur Bhargav Sosale, I developed a deep learning solution to enable every doctor to rapidly detect diabetic retinopathy, a complication of diabetes that can cause blindness if left undiagnosed.  

Fast forward five years and today, diabetes patients can be quickly and easily screened for this condition at their primary care centre. Medios is now the AI arm of Remidio, an ophthalmic devices startup. Our AI is integrated in the low-cost, portable retinal cameras we produce, and the solution is currently being commercialised in India and Europe, as well as Singapore. 

A nurse in India using the Remidio FOP device that operates on Medios’ AI.

It’s been an exciting journey, so by becoming a mentor I want to inspire others to explore the MedTech field as well. I get great personal satisfaction in giving my knowledge back to new students who are also keen to get involved in developing products that can make a real difference to people’s lives.   

I also know just how difficult it is for young MedTech startups to find the talent they need. While there’s massive demand for people skilled in developing AI- and machine learning–based solutions, we’re facing a shortage of qualified workers. By opening Medios up to apprentices, we’re hoping to help bridge that talent gap.  

At many universities, teaching students how to leverage AI is not yet a core part of the curriculum, partially because it’s such a new technology. As a result, many students interested in knowing more about AI need to look outside the scope of academia for places where they can explore applications that solve real-world problems. That’s where apprenticeship opportunities with a company like Medios comes in. 

Taking part in a six-month apprenticeship means students get to work on just those kinds of projects with the support of a specialist, such as myself. That can really help them deepen their competencies.  

It starts with a stringent interview process, where I ask the potential mentee to invest a few hours working on some difficult problems. Then, in a debrief session, we look at solutions they might not have considered. I’ll also be looking to gauge how much energy they’re putting in to solving the problems, whether they are really invested in learning more and whether they’d be motivated enough to try and figure stuff out by themselves.  

If the apprentice joins us, I give them total ownership over a project and ask them to improve on it. That might involve working on a component of the AI-based solution we’re currently building to screen for retinopathy of prematurity, a visual disability that some preterm babies develop. It’s a leading avoidable and treatable cause of childhood blindness and quite a complex disease, so we’re building several AI models to create a full solution.  

I’m always on hand to guide and help our interns, but the idea is that they first try to learn by themselves. I believe this approach can really motivate people to progress. It also helps them feel that they’ve contributed something meaningful by the end of their six-month stint. 

Florian standing in front of the image of an eye with signs of diabetic retinopathy. The image is one of thousands that goes into Medios‘ dataset when developing an AI solution.  

In terms of soft-skills and general business-skills development, the apprentice gets to participate in everyday discussions with senior people on the team. I also connect them with our medical director, who trained as a doctor and doesn’t have a technical background. That helps them develop ways of communicating clearly about their progress in non-technical language.  

We’ve had some great success stories. Recently, for example, I mentored an apprentice who had initially trained as an accountant before deciding to switch to tech. After completing a six-month immersive course in data science, he approached us and I gave him an internship opportunity. He performed at such a high level that he became a permanent employee and stayed with us for over a year.