The difficult thing about plumbing is that nobody admires it when it works. Turn the tap, get the water, move along. Maitham Dib has chosen a similar bargain for his company. Junction sits behind lab orders, device connections, status updates and streams of data. Its customer sees an API. Underneath, there are laboratories, logistics, incompatible formats and more than a few stubborn remnants of the fax age. The cleaner Junction makes the experience, the less anyone notices Junction at all.
That is a curious destination for someone trained to think about aircraft. Dib studied aerospace engineering at the University of Bath, graduating with a first-class MEng in 2016. His early work at Rolls-Royce included development engineering, data analysis, a graduate scheme and machine learning. On his public profile sits a university project with the wonderfully literal title “Utilising Smartphone Optics to Deliver a low cost Side Collision Avoidance System.” The prototype used augmented reality and a phone camera to detect dangerous lane encroachments. Even then, he was trying to make complicated sensing useful at the edge.
The work earned formal notice. Bath gave him its BP Centurion Award for academic excellence and contribution to university life. His profile also lists the Royal Aeronautical Society's Royal Aeronautical Prize and patented work involving drains for pressure-sensing lines. These are not the usual origin myths of a software founder. There is no dorm-room marketplace, no social app assembled over pizza. There are control systems, sensors and the sober possibility that an interface might have consequences.
From wings to messy data
After Rolls-Royce, Dib moved into digital health as a machine learning engineer at Babylon Health. The industry changed; the underlying puzzle did not. Data arrived from many places. Systems had to talk. Reliability mattered. A useful output depended on dozens of hidden inputs behaving themselves.
In a long-form founder interview, Dib recalled watching engineering teams wrestle with wearable integrations and at-home testing operations. Every device maker seemed to have a different API. Some used OAuth. Some did not. Laboratories added people, logistics and documents to the equation. Results could still arrive as PDFs. A product team that wanted one feature first had to become a reluctant expert in everybody else's machinery.
His initial answer was Vital, a company accepted into Y Combinator's Winter 2020 batch. The premise was narrow enough to explain at speed: one interface for collecting data from wearables and medical devices. Narrow, however, did not mean static. Customers collected the signals and then asked the obvious next question: what do we do with them?
“I have to be infinitely curious about the problem.”Maitham Dib, Founders and Builders
That question pulled the company toward lab testing. Dib and his team did not arrive there through a grand whiteboard exercise. They followed requests from early customers, including other Y Combinator companies. The move added scheduling, orders, specimen logistics, lab partners, results and customer support. In other words, it added precisely the work most software founders would prefer to leave outside the building.
Dib's language about this period is refreshingly unvarnished. Integrating with labs was “a nightmare,” he said in the interview. His pitch was not that Junction would eliminate complexity from the world. It would absorb that complexity so its customers could make an API call and get on with the product they actually meant to build. A neat abstraction is often just a mess with one very patient owner.
The roadmap was hiding in the friction
There is a founder's method inside Junction's expansion. Start with the part you understand. Put it in front of customers. Notice where their work becomes manual again. Build the adjacent layer only when it belongs to the same system.
Dib has offered his own compact version: keep speaking to customers, keep iterating and try to sell something before writing a line of code. Payment is useful evidence, he argued, provided the customer does not promptly leave. It is practical advice with a small barb in it. Enthusiasm is cheap; a recurring invoice has standards.
He once described Vital as a “Twilio for healthcare.” The comparison was less about borrowing glamour than borrowing a design principle. Twilio hid telecommunications complexity behind developer-friendly calls. Junction wanted to do that for wearable feeds and lab workflows. The company was selling shovels, as interviewer Aakash Shah put it, but these shovels came with webhooks.
The company announced a $2.9 million seed round led by Point Nine in 2022. Three years later came an $18 million Series A led by Creandum, with participation from Y Combinator, Point Nine, Amino Collective and Inflect Health. At the announcement, the platform served more than 140 organizations, supported more than 500,000 lab tests a year and connected over two million devices.
The numbers described scale, but the rebrand announced alongside them described identity. Vital became Junction. Dib said the new name reflected the company's role in connecting systems, data and patient care. The word is almost aggressively functional. A junction is valuable because of what passes through it and where those things can go next. It is a piece of the map, not the destination printed in large type.
A real endpoint, on a Friday
Founders reveal themselves in the systems they design for people, not merely machines. A Junction engineer, Krishan Wyse, published an account of the company's trial-day interview. It was remote, structured and paid. The assignment was not a theatrical puzzle. Wyse had to add a PATCH endpoint to an existing resource in the live codebase.
Dib opened the day in a 10 a.m. call. An engineer served as a buddy. The candidate toured the code, asked questions and wrote an RFC. The endpoint was intended for production, though the team held the release until Monday because it was Friday. That small detail contains an entire engineering culture: realistic work, real consequences, and enough sense not to tempt the weekend gods.
The company describes itself as remote-first, but not remote-only, operating around hubs in Eastern and Greenwich Mean Time. Workshops and team gatherings provide the face time; focused work happens where people work best. Dib's public hiring posts tend to be direct invitations to join the build, with roles spanning engineering, customer success, sales, product and technical support.
His writing history also leans technical. On Medium, he published explanations of API authentication and machine-learning approaches to standardizing data across wearables, alongside profiles of other founders. There is a consistent preference here: explain the machinery, show the interface, introduce the people doing the work.
“Try to sell something before you have it.”Maitham Dib, on testing demand before code
When invisible work becomes visible
In 2026, Dib wrote that people had told him he should market more as a founder. “Fair,” he replied. Junction was not the loudest company in its category, he said, but it had been quietly building. Then came the useful part: the platform now powers 17 million tests ordered annually.
It is a well-chosen unit. Funding measures investor conviction. Valuation measures a negotiation. Test orders measure traffic through the system. For an infrastructure company, throughput is biography. It says the invisible layer is being used and trusted repeatedly, which is rather more interesting than a founder posting pictures of a whiteboard.
Dib's aspiration remains larger than the interface. Junction's stated mission is to provide infrastructure that enables diagnosis ten years earlier at one hundredth of today's cost. The company language is ambitious; the daily work is concrete. An order has to reach the right lab. A device has to connect. A result has to arrive in a form software can use. Grand outcomes are assembled from small handoffs that refuse to break.
That may be the most transferable idea in Dib's story. Careers do not always pivot so much as reveal their recurring subject. The aerospace student building a collision-warning prototype, the Rolls-Royce machine-learning engineer and the founder integrating laboratories all share an obsession with systems at their boundaries. Components become useful when they communicate. Complexity becomes a product when somebody gives it a clean edge.
Most of us notice infrastructure only when it fails. Dib is building a company on the opposite proposition: make it work so reliably that customers can forget how strange the underlying world still is. No parade is required. The tap turns. The data arrives. Everyone moves along.