Somewhere right now, a physician is sitting between patient visits, opening up an AI coding tool, and trying to build their own healthcare workflow.
That sentence would have sounded absurd even a year ago. Now, according to Sofya founder and CEO Igor Couto, it is becoming surprisingly common.
“We are seeing a movement when physicians and health systems are using AI to program by themselves,” Couto told Refresh Miami. “Physicians are vibe coding.”
Behind the meme-worthy terminology sits a much larger shift happening across healthcare. After years of skepticism around AI in medicine, doctors are now saddling up to build themselves.
That changing attitude has created an opening for the Miami-based startup, which is building programmable clinical intelligence: AI systems designed to reason through medical decisions in real time while adapting to how specific hospitals and physicians actually practice medicine.
When Refresh Miami spoke with Couto last September, Sofya was still introducing the concept of a “medical second brain.” At the time, the idea sounded futuristic to many clinicians wary of handing over too much authority to AI. Now, the conversation has shifted.
“I think people are more used to having the AI thinking together,” said Couto [pictured above]. “And being way more than just a personal assistant.”
Healthcare’s first major AI wave focused largely on automation and documentation. Physicians quickly embraced AI scribes that saved hours of administrative work. Hospitals began adopting AI tools for scheduling, claims management, and operational workflows.
But eventually, some physicians wanted to go deeper. That shift pushed Sofya further into clinical reasoning itself, especially in complex specialties like oncology. The company recently completed a program through the Mayo Clinic platform, where it worked on AI models for precision oncology focused on prostate cancer patients.
According to Couto, Sofya structures patient data and clinical guidelines into reasoning systems designed to help physicians navigate difficult treatment decisions in real time.
The company is now deploying its technology across hospitals and clinics in Latin America while beginning early U.S. implementations. One key partner has been Hospital Israelita Albert Einstein, one of Brazil’s leading hospital systems, where Sofya has been validating its clinical workflows.
In the United States, Sofya is also in discussions with institutions including the University of Miami and hospital systems in Texas.
Still, the most interesting part of Sofya’s evolution may be what the company learned from watching doctors attempt to build AI systems on their own. Couto recalled conversations with physicians experimenting with AI coding tools to prototype custom workflows.
While he understood the appeal, many of those projects quickly ran into problems involving security, interoperability, and long-term maintenance. That realization pushed Sofya to rethink its own platform.
Now, Sofya is building what it calls a programmable clinical studio: an environment where hospitals and physicians can customize AI workflows on top of Sofya’s infrastructure without having to build everything from scratch.
The idea is that a healthcare system in Miami could design workflows tailored to its own patient population, while specialists could potentially create reasoning models deployable across clinics worldwide.
“Instead of Sofya specializing the workflow, we have a clinical studio environment,” Couto said. “Like Claude Code, but for physicians.”
For Couto, that collaborative approach also reflects why Miami has become such an important place for the company’s growth.
“When I talk about Sofya, it’s never the traditional cold way of the VCs,” he said. “People really want to see what you are building succeed from the core.”
READ MORE IN REFRESH MIAMI:
- Sofya bets on Miami to scale frontier AI in healthcare, building a second brain for doctors
- The healthcare system is bleeding money. Predictive AI is the only way to stop it.
- Papa launches ‘Papa Plus’ to transform companion care into an engine for quality healthcare outcomes
- UM students launch AIdMD, reimagining electronic health records with built-in AI
- What happens when a digital twin grows up? FIT:MATCH’s leap into sports and health
