moonai-kr/gcp-training

This repository is a hands on training course for learning Google Cloud Platform, or GCP, structured as a series of lessons that build one real web app step by step. It is meant to be forked, meaning you make your own copy on GitHub, then follow the lessons inside your own copy while adding code as you go. The course starts with a basic starter app: a Flask backend, a Python web framework, paired with a React frontend, a JavaScript framework for building interactive interfaces. From there, the first lesson covers Google's command line tool and web console for managing cloud projects. The second lesson adds real user sign in, using Google's Identity Platform and Firebase, letting users log in with email and password or with their Google account, and having the backend verify who is actually making each request. The third lesson builds a personalized chatbot using Vertex AI, Google's AI platform, specifically its Gemini models, combined with something called Memory Bank, a managed service that automatically extracts and remembers facts about a user across conversations, like preferences or things they have mentioned before, so the chatbot can personalize its replies over time. The lesson explains the difference between short lived memory, which only lasts for the current server session, and this longer term memory that persists across restarts. The fourth lesson packages the whole app into Docker containers using a multi stage build, one stage that builds the React frontend and another that runs the Python backend, aimed at producing a small, reproducible image ready to deploy. Later parts of the course, based on the project description, also cover deploying with Cloud Run, running things at scale with Kubernetes through GKE, and monitoring logs and databases. The training is designed to be done cheaply, noting it should cost no more than one dollar using free trial credits.