TransmogrifAI

Why we included this project

Teams already invested in Apache Spark and Scala are the natural audience here. TransmogrifAI automates the grunt work of building structured-data models, covering feature engineering, feature sanity checking and selection, and model selection across classifiers, so a cleaned Spark DataFrame can become a trained, evaluated pipeline without hand-tuning every stage. Its distinctive angle is compile-time type safety: features are declared and checked at build time, so data-type mistakes surface before a long training run instead of failing at runtime. That suits production pipelines where modular, reusable feature definitions matter more than chasing the latest deep-learning architectures. If you want AutoML that slots into an existing Spark cluster rather than standing up a separate platform, this is a solid choice.