This mobile NICU unit aims to improve care for premature infants

Salim Kandedi was born 17 weeks early. As a micropreemie, he had a less than 30 percent chance of survival — in his home country of Gabon, 16 per cent of babies are preterm and the neonatal mortality rate is 22 deaths for every 1,000 live births. Against all odds, he “lived and thrived,” as he puts it. And although he may not have conscious memories of those early days, they have shaped his professional mission.

While studying biomedical engineering at the University of Guelph, Kandedi threw himself into exploring how to bring health-tech innovations to market. Well before completing his degree in 2024, he worked at various medical device startups; those jobs fuelled his desire to develop a product that could improve outcomes for some of the world’s most vulnerable individuals. 

After he graduated, Kandedi founded Corvita Biomedical, a company dedicated to creating “an advanced life support medical device for premature babies.” His goal is to improve outcomes for preterm infants particularly in regions that lack the resources to support the kinds of technological interventions that can literally be the difference between life and death. Corvita’s product consists of two parts: the ARK Incubator, a portable NICU unit that includes a suite of monitoring, diagnostic and life support tools, and NOA, an AI-powered software tool that leverages computer vision and sensor inputs to provide real-time insights. 

Here, Kandedi explains how he learned about the power of innovation, how his algorithm can improve inclusivity in care and how he’s carrying on his mother’s legacy. 

Your background is in biomedical engineering. How did you wind up taking a more entrepreneurial path?

My mother is an electrical engineer and an entrepreneur; she instilled that spirit in me. I remember her saying that as an engineer, you can do so much for this world, and if the chances are not out there, you can always start out on your own. When I was 16, as a student just getting into engineering, I got my first real job at a firm that was linked to [the University of Toronto’s health innovation hub] H2i and MaRS: Shiphrah Biomedical. We were looking into how to help pregnant mothers have a better position while they were sleeping, because sleep position affects the baby. I started as a research assistant and moved up the ladder — leading the R&D side, then eventually manufacturing, quality control and launch. When one of my friends was pregnant, I gave her our product. A few months later, I held that baby, and I understood what innovation can do. It’s not just cool technology you create. There are real problems out there that have no solution, and if you sit down and do the work, you can make other people’s lives better. 

Neonatal care is a somewhat unexpected focus for a young guy with no kids. What sparked that interest?

When my mother was pregnant with me, she had a uterine wall rupture from a previous C section and ended up having to deliver me at 23 weeks. They weren’t sure if I was going to survive, but I did. Jump 23 years later, and I was working in robotics at a company called Able Innovation. A friend of mine had an aunt who was the president of an NGO and they couldn’t get incubators in remote areas. They knew I was a biomedical engineer, and everybody knew I loved building stuff at home, so they asked me to build a portable incubator for them. As I talked to people, I saw that this is not just an NGO problem — it’s also a problem in Canada’s North. If you go up north, babies don’t have the same chances. Mothers have to be separated from their families, sometimes for weeks before they have to deliver, because the hospital in their community cannot afford the equipment they need to support the baby. 

I imagine similar issues could come up in Ontario, where our hospital system is under-resourced as well.

When I talked to people from big hospitals — SickKids, Sinai, Sunnybrook — I saw that it’s very expensive for them to expand their NICUs. You need to spend $150,000 per bed to increase capacity, and on top of that, add $200,000 per month in staff. And you have to figure out some way to buy 12 devices per baby every five years. It’s quite complex. For the hospitals here in Canada, you have to make sure that they can increase capacity at a very low cost just by having one device, one transaction. You can bring the technology to the North and make it infrastructure independent. Those recommendations came straight from doctors. 

The ARK Incubator isn’t simply a single-device solution, correct? There’s also an AI component?

The most important task in a NICU is monitoring and diagnosis, because that’s what leads to intervention. We have very smart doctors who can figure things out. But if you fail at monitoring, your baby can perish. For the past year, we’ve been looking at jaundice. The transcutaneous sensors on the baby will sometimes have false readings for infants with darker skin, so infants of Black, Indian, Southeast Asian, mixed, Mexican and sometimes South American descent can have hyperbilirubinemia at a more advanced stage. So, there are more brain damage cases caused by jaundice in populations that have dark skin, and nobody’s doing anything about it. I put a few cameras in the ARK Incubator, and because I’ve connected those sensors to a central computer, I can build the first advanced AI neonatologist right into the device. It will be able to assess symptoms and screen for jaundice in all babies, regardless of skin colour, because we’re looking at all the little things that doctors would look into.

A challenge with AI is that algorithms often replicate systemic errors. Unless the models are trained on diverse data, they can default to standards that are based on a very limited demographic.

The ARK Incubator and NOA account for demographics: if your infant is of African descent, the algorithm will automatically change. We’re still in the early stages — the computer vision model is near completion, the product has been designed. Now we’re doing the hard task of building it. My plan is to pilot the device in Canada, Panama, Guyana, three countries in Africa, one in the Middle East, maybe in Asia. We want to train NOA — the software tool — on data that spans all those populations, so we can service them in an equitable way. I like representing Canada in the world, but I want to build something that’s good for everybody.  

When you first told your mother about this idea, how did she respond?

She started laughing and said, “Wow, I would have never thought that experience would’ve turned into something big.” She told me to go for it, and she supported me both financially and by putting me in touch with people she knew. My dad was also there, and this was the first time I’d heard his side of the story. He said, “There was nothing I could do. I was going to lose my wife and my son. The baby has been delivered and taken away, and my wife is still on the operating table. I didn’t know who to follow.” My dad said he saw the doctor trying to resuscitate me for a long time. It was a night shift, so after a while the doctor had to come back and work on my mom. But the nurse said, “We can’t let that baby die,” and she worked on me for a long time. My dad said, “If you can do something good for this world, make sure that parents who are feeling that kind of helplessness can trust in the fact that something will look after their babies.” New parents will feel helpless regardless, but this can give them a little more support and comfort.

Sarah Liss writes about technology for MaRS Discovery District. Hospital News has partnered with MaRS to highlight Canadian innovations in health and life sciences.