Gabriel Benedict is building what Visium has never had before

When Gabriel Benedict joined Visium, he was not hired to inherit a team. He was hired to imagine one.

His mandate was unusually open and unusually demanding: build Visium’s first R&D function from the ground up and turn a strategic ambition in life sciences into a focused research programme.  The first flagship project is exactly that. In partnership with a major pharmaceutical company and a compute partner, Gabriel and his team are developing a protein foundation model, part of a broader system that uses machine learning to identify viable proteins and molecules at a speed conventional screening could never match.

In drug discovery, the search space is almost impossibly vast. Somewhere inside it may be a molecule that binds cleanly, behaves predictably, and can one day become a therapy. The difficulty is not that there are no answers. It is that there are too many places to look.

That is the kind of problem Gabriel has spent much of his career learning how to solve.

His PhD was in generative retrieval, a field concerned with training systems to find the one result that matters inside an enormous universe of possibilities. At the time, the setting was entertainment: how to place the right film in front of the right viewer on a streaming platform. Today, the stakes are different. The object is no longer a film, but a protein. The audience is no longer a viewer, but a biological system. Yet the intellectual challenge is strikingly similar: from a near-infinite field of candidates, surface the one that matters.

A career built around search, systems, and scale

Gabriel’s path to Visium has not followed a single narrow track. It has moved between clinical research, academic machine learning, product engineering, and team building. Each chapter now bears directly on the work he is doing.

At Cytel, a contract research organisation, he worked across phase one, two, and three clinical trials. It gave him direct exposure to the mechanics of drug development: the timelines, the evidence, the constraints, and the statistical discipline required to move a therapy from hypothesis to human reality.

Later, through an industry PhD jointly held with the University of Amsterdam and RTL Netherlands, he developed the generative-retrieval expertise that now underpins his work in protein design. What began as a question about recommendation and relevance has become a question about biology, probability, and therapeutic potential.

Then came Amazon, where Gabriel built an edge-AI team from the smallest possible starting point: a Slack channel. The work involved moving AI off the cloud and onto devices, closer to where decisions were being made. He led the team until it had enough structure, momentum, and purpose for a dedicated product manager to take it forward.

That experience reveals something central about how Gabriel works. He is not only interested in building models. He builds the conditions in which models, teams, and ideas can survive.

“If projects you helped create can thrive when you leave the company,” he says, “that means you delivered outstanding work.”

It is a quiet statement, but a telling one. For Gabriel, success is not measured only by technical novelty or internal visibility. It is measured by durability. A good system should continue to function. A good team should continue to grow. A good idea should not depend forever on the person who first defended it.

Why Visium, why now?

Gabriel chose Visium because it offered a rare combination: a company focused on practical AI, an international team with deep technical talent, and close proximity to industries where the consequences of better systems are tangible.

Visium’s work sits at the intersection of advanced technology and deep industry expertise. For Gabriel, that combination mattered. He was looking for an environment where ambitious research could be translated into practical systems, and where technical exploration remained closely connected to real-world impact.

Visium offered him exactly that: the opportunity to build a research function from the ground up, backed by a clear commitment to making R&D part of the company’s next stage of growth.

The protein foundation model is only the beginning. Around it, Gabriel is shaping a team designed to think deeply, move quickly, and work across disciplines. The ambition is not simply to apply machine learning to life sciences. It is to change the pace at which promising biological candidates can be discovered, evaluated, and moved closer to impact.

The person behind the work

Outside Visium, Gabriel is a triathlete, a long-standing improv performer, and someone who speaks French, German, English, and Spanish. He credits improv with strengthening not only his language skills, but also his ability to listen carefully, adapt quickly, and respond with confidence in changing situations.

He also enjoys spending time with his family, being outdoors, and reading early-twentieth-century literature. One of his favourite quotes comes from Nicolas Bouvier:

One of his favourite lines comes from Nicolas Bouvier:

“Some think they are making a trip; in fact, it is the trip that makes you or unmakes you.”

The quote resonates with the kind of work Gabriel is doing at Visium: building something that will evolve through the process itself. He joined to help establish the company’s R&D function and contribute to its next stage of growth. The work is still at an early stage, but its direction is clear: to build research capabilities that are ambitious, practical, and designed for lasting impact. 

‍