Sumit Sharma, M.Tech.

SPIRE Lab, IISc – IISc, Bengaluru, India

• Work with dermatologists to develop a suite of image analysis and deep learning solutions to assess skin health from cell phone photos.
• Own the full machine learning stack: curate images, manage labeling, develop training pipelines, and validate clinical performance of five AI models.
• Version and (automatically) deploy models behind documented APIs.

Philips Research Bengaluru – San Diego, CA

• Led the technical development of an AI-driven brain cancer app from conception through FDA submission.
• Built scalable machine learning devops using mlflow, quiltdata, and github actions.
• Designed and developed a clean, well-documented, well-tested Python package for medical image analysis.

Healthlytix (merged with Cortechs.ai) – San Diego, CA

• Designed and trained convolutional neural networks for segmentation of prostate MRIs.
• Developed two academic prototypes into robust, production-level apps and helped guide them through FDA 510(k) clearance.
• Wrote algorithms for registration and segmentation of head CTs for use by ER radiologists.

Uptake – Chicago, IL

• Led a small Data Science R&D group responsible for building Uptake's capabilities in image analytics.
• Developed object detection and change detection algorithms using both classical image analysis and deep learning for use with satellite imagery.
• Designed, implemented, and trained convolutional neural networks; personally curated and labeled a 30,000-image dataset.

Research Computing Center (RCC) – University of Chicago, Chicago, IL

• Scripted image processing – segmentation, registration, filtering, tracking – for microscopy and MRI research.
• Installed software and provided tech support for users of RCC's 13,000 node cluster.
• Taught "Image Analysis in Python" and "Introduction to RCC" workshops.

MRI Research Center – University of Hawaii, Honolulu, HI

• Pioneered novel methods for correction of patient motion during MRI acquisition.
• Designed, implemented, and published real-time motion correction algorithms for use in clinical studies.
• Received an intramural NIH grant to adapt MRI motion correction technology for use in a neonatal population.
• Supported the lab by scripting data analysis, mentoring new hires, and organizing a journal club.