UnionML: A Microframework for Building Machine Learning Applications

Discover a solution to the common challenge of transitioning machine learning projects from research to production in this 27-minute conference talk from MLOps World: Machine Learning in Production. Learn about UnionML, an open-source microframework for building and deploying machine learning applications at scale. Explore how UnionML, created by the team behind Flyte, provides a simple, user-friendly interface for defining the essential components of ML applications, from dataset curation and sampling to model training and prediction. Understand how this framework automatically generates workflows for model tuning and deployment across various prediction scenarios, including offline, online, and streaming contexts. Gain insights from speaker Niels Bantilan, a machine learning engineer and core maintainer of Flyte, as he draws parallels between UnionML and standardized protocols like HTTP, showcasing its potential to streamline the ML development lifecycle and enhance productivity for data science and machine learning practitioners.