The useful thing about Pelu Tran's career is not that he walked away from a nearly finished medical degree. Silicon Valley has enough dropouts to fill a small, rather self-satisfied university. The useful thing is what he chose to keep: the habit of looking closely at how clinicians actually work, where attention gets lost, and why an elegant tool can die in the corridor between a demo and a daily routine.
Tran has built two companies in that corridor. The first, Augmedix, began with the gleam of Google Glass and the mundane burden of medical documentation. The second, Ferrum Health, lives deeper in the wiring. It helps health systems evaluate, deploy, monitor, and govern clinical AI applications. One company started with a device people could point at. The other is largely connective tissue. Both are bets that implementation matters as much as invention.
That preference for the difficult middle explains how a biomechanical engineering student, an almost-doctor, a product operator, and an enterprise software chief executive can all occupy the same résumé without looking like four different people.
The four-month wager
Tran grew up in New Jersey in an immigrant family and came west to Stanford, where he earned a bachelor's degree in biomechanical engineering before entering medical school. The sequence matters. Engineering taught him to see systems as loads, joints, and connections. Medical training put him inside a working environment where every extra click and administrative detour competes with a clinician's attention.
At Stanford he met Ian Shakil through courses that deliberately mixed students from different departments. Shakil later described the arrangement as “structured founder dating,” a phrase with the comic accuracy of someone who has survived both business school and cap tables. The pair knew they wanted to build in healthcare before they knew what they would build.
In 2012 they tried an early Google Glass prototype. They saw a way to let a clinician stay engaged in the room while a remote specialist handled the documentation. Shakil left his job. Tran stepped away from medical school four months before completing his MD. They co-founded Augmedix, with Tran leading product and commercial functions over the years that followed.
The timing supplied a tidy piece of founder theatre. In January 2015, Augmedix announced a $16 million financing. The next day, Google said it would shut down the consumer side of Glass. Customers understandably wondered whether the floor had disappeared. Tran had to explain that the enterprise device and the service around it were still alive.
The episode exposed the difference between a product's headline and its reason for existing. Glass was the memorable object. Documentation burden was the durable problem. Augmedix kept building, moved beyond a single wearable, and grew. Public profiles of Tran credit his product and commercial leadership during an expansion to more than 1,000 employees in five countries, roughly $20 million in annual revenue, and relationships with large American health systems.
“Every business is made up of people, and people make every decision in a business.”Pelu Tran on enterprise building
A second company goes below the surface
By 2018, Tran had spent six years watching technology companies meet the procurement habits, privacy rules, legacy systems, and crowded calendars of large healthcare organizations. He and Kenneth Ko founded Ferrum Health around a plain observation: algorithms may be plentiful, but the institutional route for testing and using them is not.
Ferrum's pitch is a platform rather than a single diagnostic application. It provides a private environment in which a health system can connect clinical data, evaluate applications against local datasets, put selected tools into existing workflows, and watch their performance over time. Tran has called the company “the last mile for healthcare AI.” It is a delivery metaphor, and a useful one. The parcel has already crossed continents. The expensive part is getting it to the correct door without losing it.
The last-mile loop
The enterprise logic is as important as the technical logic. Without a shared layer, every new application can become another integration project, security review, contract, dashboard, and support relationship. A catalog is attractive because it offers choice. Governance is necessary because choice without oversight becomes a junk drawer with an invoice.
In 2023, on Fenwick's Closing Time podcast, Tran pitched the business as a kind of “Plaid for healthcare AI,” borrowing the financial technology company's role as connective infrastructure. The analogy turns a complicated stack into a familiar shape: many applications above, many institutional systems below, one standardized route between them.
Being correct is not the same as being heard
Ferrum's early sales conversations supplied Tran with a less flattering and more valuable education. The team had data about overlooked findings and built software intended to add another check. They spoke in the vocabulary of mistakes and failure. Clinicians heard accusation. Conversations stopped.
Tran has been unusually direct about the error. The company had, in his telling, an “air of self-righteousness.” The team believed the evidence was obvious and could not understand why other people did not immediately see what it saw. Eventually, it recognized that caring, overextended professionals were being presented as the problem when the company's own thesis was that the system around them was the problem.
So the language changed. Ferrum talked about rising data volume, complicated scans, constrained teams, and the need for a dependable safety net. More clinicians listened. The product had not changed overnight. Its social interface had.
This is where Tran's interest in positioning becomes more than a founder's bookshelf detail. He says he returns to Al Ries's Positioning, the 1981 marketing book about occupying a clear place in a customer's mind. It is easy to dismiss positioning as the decorative wrapper around engineering. Tran's experience suggests something closer to interface design. Words decide whether a capable user reaches the next screen.
Trust, people, then scale
Tran's stated business principles are strikingly un-mystical. Every business is made of people. Trust is rare. Take care of employees. Find a champion and make that champion successful. They sound more like instructions pinned beside a sales desk than declarations carved into a lobby wall.
One model arrived early. During his sophomore year at Stanford, Tran worked as a corporate finance intern at Integra LifeSciences. He watched chief executive Stuart Essig and the leadership team disagree constructively, move between strategy and operational detail, and remain friends. Tran later said the experience set the bar for the culture and team he wanted to build.
The combination is revealing: an engineer's attention to connections, a clinician's respect for consequences, and a marketer's concern with how ideas arrive in another person's mind. Ferrum's platform mirrors those habits. Connect the systems. Test before trusting. Observe after deployment. Keep the institution in control.
In September 2024, Ferrum announced a $16 million Series A led by Foundry, with participation from Catalyst by Wellstar, Headwaters Ventures, and UnitedHealthcare Accelerator, alongside returning investors. The company said the round brought its total funding to $31 million. It also reported more than 250 sites using the platform and more than 2.5 million unique records analyzed by that year.
Funding is a milestone, not an ending, especially for a company selling into institutions that measure time in committees. In interviews through 2025, Tran kept returning to governance, local validation, aging infrastructure, and the hazards of assuming that a model that worked elsewhere will work identically here. The vocabulary is sober because the market has moved beyond asking whether AI can produce impressive results. The operational question is whether an organization can repeatedly decide which results to trust.
“One of the rarest resources in the business is trust.”Pelu Tran
Keep the problem, change the machinery
Look backward and Tran's two-company career forms a clean argument. Augmedix outgrew its identity as a Google Glass company because the burden it addressed survived the device. Ferrum is designed to accommodate many AI applications because no responsible health system should have to rebuild its institutional plumbing for each new model.
The common unit is not the gadget or the algorithm. It is the workflow. What does the clinician have to do? Where does the data sit? Who approves the tool? How is performance checked? Who notices when conditions change? These questions lack the romance of a launch video. They have the virtue of still being relevant on Monday morning.
Tran once exchanged a nearly completed professional credential for a company built around a then-novel pair of glasses. The risky part was visible. The durable part took longer to see. Fourteen years later, his work is less interested in spectacles, in both senses of the word. It is interested in the connective layer that lets useful technology become ordinary. There are worse ambitions than making the remarkable reliable.