Matt Hollingsworth dressed as Albert Einstein in third grade. Later, he set his sights on CERN. It is a satisfyingly orderly beginning for a career that would eventually include equestrian score sheets, telecommunications networks, connected devices and a chief executive’s chair. The costume was easier to explain than some of the job changes. Yet follow the work closely and the apparently scattered pieces begin to fit: measurements, software, and the question of what someone can do with the result.
He earned bachelor’s and master’s degrees in physics at the University of Tennessee, Knoxville. His research took him into the collaboration at CERN that discovered the Higgs boson. That achievement belongs to a large scientific team, with Hollingsworth contributing as a member of the collaboration. His contribution involved both analysis and the software needed to handle the accelerator’s data. The business career came after an apprenticeship in making large quantities of information usable.
Physics gives his résumé a memorable opening. Software gives the later chapters continuity. An experiment needs a way to record what happened, inspect it and decide which patterns deserve attention. Those tasks travel rather well. A detector and a commercial product have different purposes, but neither becomes useful merely because someone has assembled complicated machinery. Somebody still has to operate it, understand the output and trust the answer.
The instrument needed a human interface
One of Hollingsworth’s early projects makes that point unusually concrete. Between August 2011 and February 2012, he designed a web interface for the Pixel Luminosity Telescope’s control system. The instrument used sophisticated detector readout electronics. Operating the underlying control software by hand demanded substantial experience. His assignment was to make the system easier to operate, allow its state to be logged, and support checks on the data it produced.
There is a useful distinction here between making an instrument work and making it workable for someone else. A control interface sits at that boundary. It translates a specialist’s knowledge into actions another operator can carry out. The designer must preserve the complexity that matters while reducing the complexity that gets in the way. Hollingsworth was already working on that boundary before he went to business school.
His publications also include the CMS collaboration’s 2012 paper announcing observation of a new boson near 125 GeV. The formal scientific result and the practical interface project sit beside each other in his record. One concerns what the experiment established; the other concerns how people use an instrument. Together, they offer a more detailed introduction than the convenient label “physicist turned founder.”
A detour with reins attached
A conversation over coffee with fellow CERN researcher David Stickland helped carry Hollingsworth into dressage analytics. He worked on Global Dressage Analytics, a web application for riders to track progress, compare performances and receive training suggestions. A sport with horses, elaborate movements and judges might look an unlikely destination for a particle physicist. It did, however, have plenty of numbers.
The service brought detailed competition results together so users could examine riders, horses and judges. Riders could also upload results from national shows. Its proposition was straightforward: a final score leaves useful detail hidden. Looking across the components can reveal where training deserves attention. The horse still has to do the movement; the spreadsheet has yet to acquire that talent.
Global Dressage Analytics says its tools and analysis were used by all three medal-winning teams preparing for the London Olympics. That is a claim about the service’s use, rather than proof that software won anyone a medal. It nevertheless places the work in a concrete setting. The audience included people making consequential decisions about preparation, rather than people admiring a demonstration.
Stickland’s background makes the connection less mysterious. He was a Princeton research physicist working at CERN who applied analysis to dressage judging. Hollingsworth worked on the application from May 2011 to March 2014. The move into business grew out of a collaboration and a specific need. It was a small market with a recognizable user, which is a rather useful thing for an early software venture to have.
The packets came next
By December 2014, Hollingsworth was a senior software engineer at DeepField Networks. He analyzed network telemetry: the information that helps explain traffic patterns across computer networks. That month, a computing society scheduled him to speak at Atomic Object about distributed computing with Hadoop. The subject was a practical one: how to spread analytics across a cluster of ordinary machines.
The announced talk covered more than the familiar MapReduce model. It included stream processing, distributed machine learning and service coordination. These are the sorts of subjects that make a software system behave under load. They tend to receive less applause than a striking result, and considerably more attention when something stops working. His work had moved from an accelerator’s data to the traffic of large networks.
He was also the technical reviewer for Learning Predictive Analytics with Python. Reviewing a programming book is a quieter entry on a résumé than participating in a scientific discovery. It is still a useful clue about the kind of work he was doing: checking explanations and implementations that other people would rely on. Across the research, engineering and review work, the practical details remained visible.
Learning to carry the commercial side
Hollingsworth also proposed, won and managed projects for the Department of Defense. The verbs matter. Winning a project asks for more than an interesting technical proposition. Managing it adds responsibility for delivery. His career also included artificial intelligence work at Soar Technology, bringing the engineering thread into another organizational setting.
He then undertook product management for Internet of Things applications at Samsung. A connected device brings its own collection of questions about what the user needs, how the software behaves and which features belong in the product. His background now included research, a startup, technical leadership and product management inside a large company. The scale changed; the need to explain what the technology was for persisted.
Stanford’s Graduate School of Business supplied the next part of that education. He studied there from 2015 to 2017 and earned an MBA. His activities included the Big Ideas Club and the High Tech Club. The names are almost a miniature portrait of the curriculum he was assembling: room for ideas, followed by the rather less forgiving business of deciding which ones can support a company.
The company he had not set out to start
Carta Healthcare was co-founded in 2017. Hollingsworth became its chief executive. He has described the beginning with a disarming sentence: “I wasn’t really setting out to create a new company.” The work started in research at Stanford and developed into a venture. That sequence matters to the story of his career. A business emerged from work he was already doing, rather than arriving first as a title he wished to hold.
The chief executive role lasted through the company’s early development. In December 2023, Carta announced Brent Dover as his successor. Hollingsworth remained a board member, with a continuing focus on product innovation, business development and strategic partnerships. Those responsibilities bring together the strands already present in his previous work: making something, finding its commercial purpose and working with people beyond the engineering team.
The appointment arrived during a period of expansion. Carta reported threefold revenue growth in 2023, a 50 percent increase in its organizational customer base and a fourth consecutive year of complete customer retention. It also reported a $25 million Series B round. These are company figures for that period, rather than personal trophies. They provide context for the handover: an enterprise moving into another stage, with its founder taking a different seat at the table.
Founding and running a company are often compressed into one verb. His transition makes the separate responsibilities visible. The board role preserved a connection to the business he had helped create, while another executive took on the top operating position. A career with several changes of field had acquired a change of perspective inside the same organization.
A useful person to have in the room
In 2023, Hollingsworth joined the Oregon Bioscience Incubator’s mentoring program. Asked what colleagues appreciated about him, he described his ability to connect technical capabilities with product and sales strategy. His warning to engineers was compact: “cool technology cannot always be sold.” His warning to sales teams went in the other direction: a saleable idea may be something that should not, or cannot, be built.
That is a pleasantly inconvenient job description. The engineer has a prototype; the salesperson has a prospect; someone has to establish whether both belong in the same business. Hollingsworth described helping teams find product-market fit, especially in complex industries where familiar startup methods can struggle. It is a self-description grounded in the work he had done, from technical projects to commercial leadership.
Mentoring gives that experience another audience. An early founder can need someone who understands the excitement of getting software to work and the frustration of discovering that a working system has no obvious buyer. His account acknowledges both. The commercial question comes with technical limits, and technical enthusiasm comes with commercial limits. The person who keeps both in view may occasionally spoil a meeting. He may also make the next meeting more useful.
“cool technology cannot always be sold”
Matt Hollingsworth, on connecting engineering with sales
Another beginning, with more experience behind it
In March 2025, Hollingsworth publicly announced a new role as a co-founding partner at Recombinate Health. He expressed gratitude to the Carta team and described wanting to apply what he had learned to technology that made people’s work easier. The announcement carried the tone of a new beginning while retaining an explicit connection to the colleagues and company that preceded it.
His newer public affiliations extend further. Kyndred lists him as a co-founder alongside Matt Survis, with Johannon Olson as a partner. He and Survis are also affiliated with Mimi & Grace, a business founded in 2026. The career has continued beyond the former CEO title, even as Carta remains part of his professional identity.
By March 2026, he was articulating a further ambition: better jobs through more deliberate adoption of AI. He argued for designing technology around the people whose work it changes, including what they enjoy doing and how they can develop. Training and employee participation in product design were part of his account. The emphasis puts another responsibility on a founder. Time saved by software presents a choice about what people do next. His preferred approach invests that capacity in their work and growth. It brings the person using the system into the discussion early, while there is still an opportunity to shape it.
There is no tidy industry label that captures the whole route. Physics degrees, an instrument interface, riding scores, network analytics, product management and company leadership belong to different worlds. Put the work in order, though, and a consistent question emerges. Who will use this, and what will it let them do? Hollingsworth has repeatedly worked where those questions meet the software. His record makes room for the clever idea and the effort required to make it useful.