The Young Designer Bringing Old-School Craft to AI Software

Good software design is crucial to making sure people can properly use and navigate digital services. The strongest products aren't the ones that look the most advanced, but the ones that make complexity feel simple.

Paul Faivret, a 24-year-old French designer now joining New York's startup scene, shows what that idea can look like in practice. From building small websites in rural France to shaping AI platforms used by professionals, his work has centered on making software feel straightforward and intuitive. Now at Interfere, a company developing a product that finds bugs in software, Paul continues that approach, aiming to provide designs that are reliable, understandable, and easy to use.

A Design Foundation At A Young Age

Paul's interest in design originally started as an interest in photography. As a teenager in rural France, he took up photography but, as he pursued it, realised he was more interested in building a website to show his images than in taking the pictures themselves. "I wanted a portfolio," he says. "No one visited except my mom, but I loved building it and kept rebuilding it every two weeks instead of doing photography."

That habit became his first training ground. He redesigned his mother's running club website, then others for local groups, and by fifteen had started a small web design business — as he recalls, one of the youngest in his region at the time.

Those early projects shaped how he conceived of design and what he prioritises in every project he's come across since. "Clarity and usability matter most," he says, realising the importance of aspects like font selection, spacing, and kerning — little details that make a design feel (or look) effortless.

His first professional test came at Metacell, a U.S. company focused on building software for the life sciences. There, he worked with scientists whose tools were complex and whose priorities left little room for design, requiring him to apply his skills in a brand new context. "Life sciences didn't really care about design," he recalls. "But people still need to have good, clean software."

The job taught him to understand how design decisions can directly correlate to how software works, with every interface decision affecting researchers' ability to work. Design, he realized, could be used as a way to make specialized work easier to understand.

Breaking the Mold at V7 Go

When Paul joined AI-driven startup V7, the company was best known for its computer vision platform. The team later shifted focus to a new product, V7 Go, a tool that provides summaries or large documents for people working in fields like insurance and finance.

As the only designer on the project, Paul worked in an industry that was rapidly adopting chat-style interfaces after the launch of ChatGPT. Feeling that format could have serious limits for professional use, he, instead, created something different: a structured dashboard that let users organise and analyse large volumes of information at once.

The interface allowed users to upload hundreds of contracts or insurance files and add columns driven by prompts such as "flag missing clauses" or "summarise risk level." Each cell produced a structured response, creating a live, searchable dataset instead of a chatbot's typical stream of conversation, keeping the familiarity of a regular spreadsheet.

"While everyone went the way of chatbots, where they would drop documents and ask a question, we went the other route and solved the problem at scale instead of one‑by‑one," he recalls. That approach helped V7 Go stand out at a time when most AI products were still looking for how to make the technology fit real work.

Designing Self-Healing Software at Interfere

This view on AI eventually led Paul to Interfere, a New York–based startup developing a product that aims to spot, categorise, and patch bugs in real-time.

At the time, he was working at Permanent, a design studio that partners with early-stage companies to build their product foundations. When Interfere reached out, the concept immediately appealed to him: a system built to keep digital programs reliable without repeated human intervention.

As founding designer, Paul is responsible for how the product looks, behaves, and communicates results to engineers. His job consists of building and refining the interface where automated bug reports become readable insights, with his decisions about information hierarchy, layout, and determining how users trust what the system finds.

For Paul, that idea isn't just technical; it's about letting users trust that the systems they use will be reliable and easy to follow. Now relocating from France to New York, he sees the role as a chance to apply the precision of European design within the speed of an American startup.

Keeping Craft at the Center of Automation

Paul conceives of design as a process of reduction: removing distractions until only what's useful remains. He rejects what he calls "over-processed aesthetics," which is the tendency toward visuals that look modern but function poorly. His work favors the opposite: his goal, he says, is to make technology feel simple without making it simplistic.

That same outlook shapes how he thinks about AI's future, as he imagines software that can, if well built, continually maintain itself. For him, autonomy in design isn't about displacing human work but preserving quality over time, through building systems that stay clear and usable on their own. "If I can help bring that philosophy to all software companies, that would be amazing," he says.

Paul Faivret's path reflects a consistent interest in making technology easier to use. His work in product design aims to show that the most lasting and accessible work in AI will come from those creating technology focused on maintaining trust, making sure that what's built keeps working, quietly and well.