An Introduction to MLOps in the Healthcare Industry Good MLOps (machine learning operations) are best practices and technologies meant to improve data science workflows. Data scientists often lack the tools necessary to put models into production, and the friction between engineering and data teams dramatically slows time to production. Like DevOps, MLOps aims to make machine learning reproducible, reusable and scalable so that data scientists can rapidly iterate and deliver enterprise value at an exponentially faster pace. I would like to discuss the current and future state of MLOps alongside my team’s approach and experience. And finally, how these approaches and tools can help quickly deliver value within an organization.
Kyle Gallatin is a former biologist turned Machine Learning Engineer for the pharmaceutical company, Pfizer. He specializes in Python programming, natural language processing (NLP) and the engineering of scalable machine learning solutions within an enterprise ecosystem. In his spare time, Kyle also mentors future data scientists at NYC Data Science Academy, frequently writes data science articles on Medium, and is a volunteer teaching assistant for high school computer science.