How do you take a breakthrough in a lab and make it available to a doctor - anywhere in the world, in real time? You turn AI into infrastructure. Within Roche Digital Technology (RDT), Deepak and his team are not just building models - they’re building the intelligence layer that connects discovery to patient care. Their work ensures that advanced insights don’t stay in the lab, but are scaled, secured, and delivered precisely when and where they can change a life by applying AI at scale.
When I first joined, I noticed that many diagnostic challenges were considered too complex to scale globally - not because we lacked models, but because we lacked the infrastructure to operationalize them. In AI, building a high-performing model is only part of the solution. We hit a wall when we realized the real challenge was everything around it: how to handle large, sensitive datasets securely, how to ensure consistency across environments, and how to deliver results fast enough to be useful in a clinical setting.
That’s where our focus shifted - from models to systems. We started building the foundations that allow AI to move from experimentation into real-world use: robust data pipelines, scalable architectures, and deployment frameworks designed for reliability and trust.
Because in healthcare, AI only matters if it works - securely, consistently, and at the moment it’s needed.
“At a tech company, you're building for a user. At Roche, I realized I’m building for a patient. When you realize your code is the difference between someone getting a diagnosis today or next week, the work becomes very personal.”
slide 1 of 2 I've learned that people will forget what you said, people will forget what you did, but people will never forget how you made them feel. Maya Angelou
"Roche has more than 30 years of experience developing innovative treatments for respiratory diseases. We are committed to applying our expertise to understand the immune mechanisms that drive COPD, so that we can bring positive change for physicians, their patients, and health care systems."
Larry Tsai Senior Vice President and Global Head of Immunology, Product Development
slide 1 of 2 I've learned that people will forget what you said, people will forget what you did, but people will never forget how you made them feel. Maya Angelou
Operating at this scale isn’t about building better models - it’s about making them work under real-world constraints. My team and I tackle high-stakes architectural puzzles every day:
Scaling AI into production: Using MLOps and containerization to turn research models into deployable services—accelerating the path from experimentation to real-world use.
Engineering for scale and trust: Designing hybrid cloud architectures that can process large, complex datasets while meeting strict regulatory and security requirements.
Advancing applied GenAI: Leveraging generative AI to structure and interpret unstructured medical data—unlocking insights that were previously inaccessible in clinical workflows.
For me, this was the turning point. It wasn't about "optimizing for speed" anymore: it was about optimizing for certainty. Every millisecond of latency my team shaved off wasn't just a technical win - it was time handed back to a patient waiting for answers.
By rethinking our approach and treating our AI infrastructure as a living, breathing engine, we transformed that "wall" into a bridge. The resolution isn't found in a successful deployment or a clean line of code - it’s found in the broader implications. We are now solving problems that were considered impossible just two years ago. Knowing that the infrastructure I help build today will be the foundation for how diseases are treated five years from now? That’s the kind of legacy you just don't find in standard tech roles.
Deepak Sondur Head of Applied AI, Roche Digital Technology
seperator
At Roche, we believe the most direct route to your goals isn't always a straight line. If your current growth is on 'pause,' it might be time for a strategic detour. Whether it’s gaining international leadership or finding a better lifestyle, your next dimension is waiting in Budapest.
seperator
For tech experts with a world of choices, Deepak’s message is simple: You can use your skills to make people click on ads, or you can use them to ensure a person gets the right treatment at the right time. At Roche, your skills don't just solve problems - they change the trajectory of lives
