Build practical skills in BigQuery Machine Learning (BigQuery ML) to develop predictive models, forecast outcomes, and work with generative AI. In this Specialization, you'll learn how to use SQL and Google Cloud tools to train, evaluate, and deploy machine learning models without moving data out of BigQuery.
You'll begin with BigQuery ML fundamentals, including model types, inference, and model management. Next, you'll explore machine learning in data pipelines, build classification models to predict visitor purchases, and use taxi trip data for fare prediction. You'll also create an expected goals model using soccer data and deploy a customer churn classifier to Vertex AI for online predictions.
As you progress, you'll explore time-series forecasting with Vertex AI before using SQL and large language models (LLMs) for text summarization. Finally, you'll work with Gemini models, embeddings, vector search, and retrieval-augmented generation (RAG) in BigQuery.
By the end, you'll be able to:
Train, evaluate, and use predictive models in BigQuery ML.
Apply classification, forecasting, and model deployment techniques.
Use LLMs and Gemini models for generative AI tasks.
Build retrieval workflows using embeddings, vector search, and RAG.
This Specialization is designed for data analysts, developers, and cloud practitioners seeking practical experience with machine learning and generative AI on Google Cloud.
Projet d'apprentissage appliqué
Throughout this Specialization, you'll apply machine learning concepts through hands-on Google Cloud labs and practical modeling activities. You'll create and evaluate models in BigQuery ML, predict visitor purchases with logistic regression, estimate taxi fares, and analyze soccer shot data using an expected goals model. You'll also train and tune an XGBoost customer churn classifier, generate batch predictions, and deploy the model to Vertex AI for online predictions. As you progress, you'll build a time-series forecasting solution using retail data, summarize text with SQL and LLMs, and work with Gemini models through SQL and Python notebooks. Finally, you'll generate embeddings, perform vector search, and create a RAG pipeline in BigQuery to support more grounded AI-generated responses.
