Isha Sharma

Talk Title:

From Model to Market: What It Takes to Launch AI in Regulated Healthcare

Abstract:

A working AI model is not yet a medical product. Between a model that performs well on test data and patients using it safely sits the launch gap: validation with real users, documented traceability, regulatory review, a quality system, cybersecurity and monitoring after release. Each step belongs to a different team, so crossing that gap is as much a coordination problem as a technical one.

The talk sets out where AI already sits in medical technology, with more than 1,400 AI-enabled devices authorised by the US FDA through 2025, and what changes when a product uses AI: bias, drift, over-reliance, explainability, privacy and accountability. It covers the requirements regulated health software must meet, including design controls under ISO 13485 and the distinction between verification and validation, alongside CDSCO's medical device software guidance finalised in July 2026. It examines how products that keep changing after launch are governed through a Predetermined Change Control Plan, illustrated with a public case study of Medtronic's AccuRhythm AI.

The second half turns to the people problem: aligning teams you do not manage, what each function on a launch actually values, and using shared data to settle conflicting priorities, drawn from a launch I ran. It closes with open questions for researchers.

Profile:

Isha Sharma is a program manager at MiniMed, formerly Medtronic Diabetes, where she leads development and global launch of patient-facing digital health apps used by patients across 120 countries. Her work covers the full product lifecycle in a regulated medical device environment, coordinating engineering, software, regulatory, clinical and quality teams from development through release.

Her background spans research, consulting and industry. She trained in research laboratories at Rajiv Gandhi Cancer Institute, CSIR Institute of Genomics and Integrative Biology, and India's Defence Research and Development Organisation. She worked as an analyst at Bain & Company on healthcare and pharmaceutical engagements, and led business development at Otomagnetics, an early-stage drug delivery startup. Her graduate research at UC Berkeley produced a U.S. provisional patent for a non-invasive osteoarthritis therapy.

She has published seven peer-reviewed papers with IEEE, Elsevier and Springer, and completed more than 20 manuscript reviews for journals including IEEE Transactions on Neural Networks and Learning Systems, Journal of Translational Medicine, PLOS One, BMC and Frontiers in Medicine. She serves as a grant reviewer for the FDA Pediatric Device Consortia program and as an expert evaluation panelist for the HHS TOPx Tech Sprint.

Isha holds an M.Eng in Bioengineering from UC Berkeley, an MBA, and a B.Tech in Biotechnology from Delhi Technological University. She is an IEEE Senior Member.