How I’m building a full Cognitive AutoML/MLOps platform in Streamlit — 305 files, 150k+ lines, audit contracts, scientific integrity and an AI assistant

For those interested in the engineering side, here is a current module-level LOC snapshot. The project has grown into a large Streamlit codebase with dedicated modules for the core pipeline, scientific integrity checks, pages, plots, E.M.I.L.I.A., meta-learning, knowledge/GraphRAG components and tests. This is one of the reasons I wanted to share the project with the Streamlit community: it is not a small demo app, but an attempt to build a full production-oriented AutoML/MLOps platform using Streamlit as the application layer.