The #Data Scientist# Who Stayed for the Problem, Not the Platform
Sandeep started in software engineering: clean code, solid architecture, and the discipline of building systems that hold under load.
That foundation matters because data science without engineering rigor is just a notebook no one can ship.
He moved into data as the discipline matured. Not because it was fashionable, but because the questions that interested him lived in the data layer: how do you predict what a patient, payer, or population will do next?
The shift from engineering to analytics to forecasting was not a pivot. It was a narrowing in on the right problem.
Persistent Systems gave him scale: complex pipelines, distributed infrastructure, and the kind of data work that teaches you what breaks at volume.
OurOffice gave him speed and ownership: a leaner environment where he owned the data stack end to end and made calls without a committee.
Both shaped how he works now: comfortable with complexity and allergic to over-engineering.
He built large-dataset forecasting models before machine learning became a product category. He architected data pipelines feeding clinical workflows before GenAI was in the vocabulary.
That early work, unglamorous and iterative and grounded in production reality, is why his healthcare AI thinking is built on what actually holds, not what demos well.
That is what Sandeep brings to Mindbowser. When a healthcare organization needs to understand how AI fits into its clinical or operational workflows, Sandeep brings a lens grounded in building and shipping real models, not theory.