The Question
"What happens when you train your mind to save lives, then discover you can amplify that mission through artificial intelligence?"
Two Languages, One Mission
I was in medical school. One day I decided to leave it and go all in on artificial intelligence. Not because I stopped caring about healing, but because I realized that what's coming with AI is something else entirely—and that I could extend that same mission from another angle.
Doctors speak in symptoms, diagnoses, and treatment protocols. Engineers speak in algorithms, data structures, and system architectures. I ended up learning to be bilingual in healing and building.
This isn't a story about choosing between medicine and technology. It's about why I chose the second language without letting go of the first: both serve the same mission—alleviating human suffering.
The Paradox of Modern Medicine
We live in the most medically advanced era in human history, yet healthcare professionals are drowning in information while patients wait for answers.
The tools to heal are more powerful than ever, but the systems to deliver that healing are overwhelmed.
The Information Avalanche
Every day in medical school, I'm reminded of a humbling truth: Medicine is the only field where ignorance can be fatal, yet complete knowledge is impossible.
The Scale: thousands of new papers a day, tens of thousands of diseases, drugs, and interactions.
The Reality: consultations that last minutes, hours lost to paperwork, and a long lag between what gets published and what reaches practice. Many decisions get made with incomplete information—not because it's missing, but because it doesn't fit into the minute when it's needed.
This is where my journey began—not with a fascination for technology, but with a profound frustration at the gap between what we could know and what we can know in the moment when it matters most.
The Awakening: When Frustration Meets Possibility
The Moment of Clarity
It was during my first clinical rotation, watching a resident spend forty minutes cross-checking drug interactions for a complex patient by hand while the backlog piled up. I thought: "What if a solid system could do this in seconds, and those forty minutes went back to the patient?" It wasn't a product in my head yet—it was an intuition: the bottleneck wasn't knowledge, it was access to it in the exact moment it mattered.
The Philosophy Shift
That's when I understood something about myself: I love to build. I love the creative part—solving, automating what today eats hours. I respect being a doctor—I know what it takes, I know how much of your life it demands—but this lights me up more, because with AI I can automate a huge amount and multiply the reach. And I didn't leave healthcare behind: I help it from another side. Technology doesn't compete with human touch. When a doctor isn't buried in paperwork and searches, there's time left for the things only a human can do: listen, accompany, and decide with judgment. A well-built tool doesn't replace that closeness—it protects it.
The Skills Transfer
Medical training had already taught me the most important AI skill: pattern recognition under uncertainty. Diagnosis is essentially a classification problem with life-or-death stakes. Building AI systems is just extending that same logical framework.
My Technical Evolution: From Scripts to Systems
My journey into AI wasn't a dramatic pivot—it was an organic evolution, each step solving a real problem I'd encountered in healthcare.
Literature Overload
Staying current with the literature full-time is impossible without help. Summaries, alerts, and automations—always under human judgment, never instead of it.
Patient Education Gap
Patients left consultations confused. Developed chatbots explaining medical conditions in plain language, available 24/7.
Clinical Decision Support
Doctors lacked access to latest guidelines. Built RAG systems retrieving and synthesizing clinical protocols for specific scenarios.
Complex Workflow Coordination
Healthcare runs on many hands. A multi-agent system can mirror how a team works, but it never executes a critical step alone: there's always a human approval and a protocol in the loop before it acts.
The Unique Advantage: Why Medical Background Matters
The Translation Layer
Most AI engineers building healthcare solutions are trying to learn medicine while building. Most doctors trying to implement AI are learning to code while practicing.
Having both perspectives means I can build bridges instead of walls.
What Medicine Teaches AI Engineering
- Systems thinking: the body teaches you nothing moves in isolation; you learn to read interacting parts, not comfortable averages.
- Risk with real consequences: in the clinic a mistake isn't a ticket, which forces you to weigh probability against harm before trusting any output.
- The fragile user at the center: you design for the person on their worst day, not for the pretty demo that impresses in a meeting.
What AI Engineering Brings to Medicine
- Scale on the repetitive and bounded: the cognitive tasks that eat hours without adding judgment are exactly where a well-governed machine frees up time.
- Decision support, never replacement: the system brings context and cites it; the person who decides and signs is still the clinician.
- Ordering the workflow, not adding screens: good AI removes friction and connects scattered signal—it doesn't hand you one more tool to tend.
Building Gnosix: Theory Meets Reality
Founding Gnosix wasn't about building another AI company. It came from the same decision: if I truly believe AI changes the game, the honest thing is to build with it and help others integrate it. Gnosix exists for that—to help people and businesses set up AI with judgment and get ready for what's coming, including AGI. The most honest way to prove the thesis was to build a system myself.
The first was Depadoc, and it's mine in the most personal sense: it's where I brought my two passions together. I loved being able to integrate the new thing—AI—with medicine, which never stopped moving me. It's a tool that helps doctors prepare for the ENARM (Mexico's national medical residency exam), and its core is a RAG system scoped to citable sources: instead of letting the model hallucinate, every answer is anchored to a medical corpus organized into fourteen categories, retrieved by embeddings and returned with the source in view. The rule I'm obsessed with is simple: if it can't cite where something came from, it shouldn't say it.
That design decision—scoping knowledge to what's verifiable, in a domain where being wrong is expensive—is the same one I learned in the clinic: sounding confident isn't enough, you have to be able to show the why. Building Depadoc was translating that medical reflex into code.
It still surprises me how much of the clinic leaked into the product decisions: every time I hesitate over showing something without a source, the doctor I used to be wins. My family didn't love this decision, and I get it. But you make choices in life knowing they're the right ones, and you keep making sacrifices—I don't experience them as loss, quite the opposite: along the way I also started the entrepreneurship side and discovered just how much AI can help you in your own business.
The Hard Truths: Challenges at the Intersection
Error has no margin: a consumer app can afford an "acceptable" failure rate; healthcare can't. That demands serious validation, not just an A/B test that looks good on a dashboard.
Regulation, privacy, and approval: long roads and a high bar. Frustrating, yes—but exactly how it should be when someone's body is on the line.
Trust is earned slowly: clinicians have already watched promising technologies fail. Credibility is built slowly and lost all at once.
Integrating with legacy systems and fragmented data: a real headache, not a marketing line. A good chunk of the work lives there, not in the model.
Practical Wisdom: For Those Following This Path
For Medical Professionals → AI
Start Small: Automate your own pain points first.
Learn by Doing: Build simple tools before complex frameworks.
Stay Medical: Your domain expertise is your superpower—don't lose it.
For AI Engineers → Healthcare
Shadow Healthcare Workers: Understand real workflows.
Learn Medical Language: Healthcare has its own vocabulary and logic.
Build with Doctors: Involve healthcare professionals in every design decision.
The Future We're Building Together
The convergence of medicine and AI isn't a passing trend—it's one more piece of how we're going to organize care—and almost everything else—at scale, with our eyes wide open to both risk and benefit. Depadoc is only the first system; with Gnosix, what comes next is carrying that same rigor—bounded, citable knowledge with a human in the approval loop—into more clinical and operational workflows, and helping others integrate AI before AGI catches us off guard. One real problem at a time.
The Vision: AI-Augmented Healthcare
- Personalized Medicine: AI that understands each patient's unique genetic, environmental, and lifestyle factors
- Predictive Care: Systems that identify health risks before symptoms appear
- Global Knowledge Sharing: AI that democratizes access to world-class medical expertise
- Human-Centered Technology: Tools that enhance rather than replace the doctor-patient relationship
The Deeper Truth
This journey taught me that the best technology is invisible—it solves the problem so well that people forget it was ever there. The best medical AI won't replace clinicians; it gives them back what the system takes away: time, focus, and room to be human in front of the patient again.
Medicine taught me to care about the outcome, not the number. Engineering taught me to scale without losing the thread. Together they taught me that the highest purpose of intelligence—artificial or otherwise—is to reduce suffering and amplify healing.
The future of care will be built by those who speak both languages: the language of healing and the language of building. I'm grateful to be, in part, bilingual in that—and I'm still building at that edge.