subtopic

Prepare the two models

Get the two things you will serve: (a) a saved ML pipeline (preprocessing + model) — your German Credit model from the ML module, or train a small scikit-learn pipeline on any tabular dataset in 10 lines and save it with joblib; (b) an LLM function answer(question) — your policy Q&A RAG from the LLM module, or a simple LLM call with a system prompt. Put both in a src/ folder with a requirements file.

Builds on: S13 · Build: Save, Pipeline & Local Deployment, Module 4 · Project 2: Company Policy Q&A Assistant (RAG — Retrieval-Augmented Generation)