If you have walked into an Aldi or a Tesco recently and picked up a bottle of water without going through a checkout, cameras in the ceiling were quietly reconciling what your bill should be. It's a world Thomas Knox knows well. He helped put ambient sensing into more than a hundred autonomous stores, rising to head of product at the San Francisco computer-vision company AiFi.
VitVio, which Knox founded and now runs as CEO out of London, does something similar to what AiFi does. Cameras and sensors in the ceiling. Models tracking who and what are in the room. Only, this time, the equipment goes into the operating room instead of supermarkets.
VitVio has a publicly announced clinical collaboration with the Royal Orthopaedic Hospital in Birmingham, and Thomas Knox says the company has expanded to hospitals across the UK, US and Europe. Thomas Knox and VitVio’s chief scientist, Grzegorz Jacenków, sat down with The Infinite Loop to talk about how their system aims to simplify the lives of staff and how the tech is far from the hardest part when working with hospitals as your customers.
TIL: What’s actually happening when your product is in action?
Knox: At VitVio, we build ambient sensing and computer-vision tech. We go into the operating room, deploy cameras and sensors, and track who’s doing what, when, where, and why. Then we use that to take over administrative work and drive efficiency across the OR block. What we’re really doing is understanding the context: the stage of a surgery, patient arrival, induction, knife to skin, wound closure, writing those events straight to the electronic health record. And because we already have a deep 3D understanding of the room, we can recreate the surgery in 4D and replay it. Put on VR goggles, stand next to the table like you were there. The goal is for VitVio to feel like another staff member in the room. One without a physical presence, that you can talk to in natural language.
Jacenków: Instead of running models that just look at the scene, we recreate what’s happening in 3D. We know where every person is, where every tool is, and we plug new [specialized] models onto that representation. If a patient has arrived, we can notify the next team to prepare the following one. If a surgery is close to finishing, we bring cleaning staff in early. We are a native 3D company.
“If you add another step to my process, you’re going to end up on the table yourself”
TIL: Hospital ROI and reduced burden on the OR team can be in tension. VitVio is pitching both. So, what’s the actual goal?
Knox: Both, and they have to be equally balanced. Everything we build has two objectives. One, make the staff’s life easier. Two, deliver hard-dollar ROI to the hospital. If you make the hospital more money but the staff’s life gets harder, you are harming the hospital in the long run. Staff burn out. Throughput drops. When we were first starting the company, we were talking to a surgeon, and he said something that became a motto: “If you add another step to my process, you’re going to end up on the table yourself.” Every part of our system has to eliminate an existing step or translate it into something easier. If we are adding a step, it’s a no-go.
TIL: What has been the hardest thing to overcome?
Knox: The tech is not the hard part. I have been part of three unicorns. Our team has come through big companies. Greg was at Amazon. I ran product at AiFi. If you walk into the O2 Arena in London, that is our tech. It runs at Aldi, Tesco, Dollar General, Marks & Spencer. The tech, we understand. The biggest challenge, and it is not just us, is that hospitals are cautious animals. Rightfully so. They are dealing with our health. But what we are proposing is not a step function; it is a leap in how they operate. It changes the entire way a hospital runs. Only the best-of-the-best fully understand where it is going. Everyone else follows, but cautiously. You have to bring them somewhere it is already working before they believe the future is real.
Jacenków: We are also privacy-first. When cameras appear in an OR, people can be afraid of what’s happening. The moment a feed leaves the OR, everything is anonymized. We do not care who is doing what on a per-person basis. And our servers are installed on-site. Data stays in the hospital. We send them notifications about the next step in the process. That is it.
VitVio’s chief scientist, Grzegorz Jacenków. Credits: VitVio
Knox: Previous versions of this technology focused on individual metrics, and that gave a Big Brother feeling. We do not track anyone’s bathroom breaks. We track the process. But we have to say that upfront. Because if I don’t tell you anything except that there are cameras and sensors in the room, your brain runs off with ideas about how it could be used against you.
TIL: Walk us through the first deployment. How was it?
Knox: Simpler than people imagine. We come in for a site visit, take out an iPhone, and do a LiDAR scan of the operating room. Less than two or three minutes. The system spits out a camera plan with the mathematical overlay of coverage. We hand it to the hospital’s team to install [the equipment] with the exact measurements. The whole install takes about four hours, overnight. The cameras aim and calibrate themselves. Everything else runs off-site. Our first deployment was up in four or five hours. It hasn’t been touched since. Compare that to the autonomous stores, where we had to redeploy 30 times because things break in the chain and it is not obvious where. Here, we had the experience already.
And we're moving into the territory now where the hospitals will deploy it completely themselves. They don't actually need us involved. So, the very first one, we'll come in, we'll white-glove it, we'll work with the team, we'll show them exactly how to do this right. But after that, we literally hand them a mobile application for an iPhone so they can do the LiDAR scan themselves, and everything else is just done through the app. All they have to do when the server arrives is just plug it in.
TIL: Is hospital self-deployment just a vision, or is it near-term?
Knox: It is already in the process. When you deploy hardware, you are almost linearly scalable. Your deployment team becomes the biggest team in the company. We wanted to grow without a two-thousand-person deployment team. So we automated it for ourselves first. Now we're packaging our systems into a tool that we're handing to the hospitals within the next 6 to 12 months or so.
TIL: You've picked up second place in Digital Health at the Nebius AI Discovery Awards. What’s your plan with the credits?
Jacenków: We are incredibly grateful to have been selected for second place in the Digital Health category, particularly from a global field of nearly 650 highly competitive submissions. We already develop specialized models designed to understand the operating room and its workflows. The Nebius credits will allow us to step up the scale, pace and ambition of that research. Our ambition is not simply to adopt advances made elsewhere, but increasingly to contribute advances of our own.
Disclosure:
This interview took place on July 1, 2026 at the Nebius AI Discovery Awards ceremony in London. The Infinite Loop is Nebius' editorial project. Editorial decisions are made independently.
Top photo: VitVio’s CEO Thomas Knox. Credits: VitVio.
