Visão Geral

O curso MLOps on Google Cloud oferece uma formação completa sobre Machine Learning e MLOps utilizando os serviços da Google Cloud Platform. O participante aprenderá desde os fundamentos de Machine Learning até a construção de pipelines de dados, feature engineering, treinamento, validação, implantação e orquestração de modelos em produção, utilizando ferramentas como Dataflow, BigQuery ML, AutoML e Vertex AI.

Objetivo

Após realizar este Curso MLOps on Google Cloud, você será capaz de:

• Simplifique e automatize as operações de aprendizado de máquina usando as ferramentas do Google Cloud.
• Aprenda o básico de Aprendizado de Máquina
• Implemente as etapas necessárias para construir um pipeline de aprendizado de máquina (ML) de produção.
• Crie pipelines de dados em lote e de streaming com o Google Cloud Dataflow.
• Realize engenharia de recursos usando Dataflow e BigQuery.
• Treine e valide modelos de aprendizado de máquina usando BigQuery ML, AutoML e TensorFlow.
• Implante modelos usando a automação do Google Cloud.
• Utilize a IA da Vertex para gerenciar conjuntos de dados de aprendizado de máquina, modelos, experimentos e implantação.
• Programe pipelines da Vertex AI usando o Kubeflow.

Publico Alvo

• Cientistas de dados
• Engenheiros de Machine Learning
• Engenheiros de dados
• Arquitetos de soluções em cloud
• Profissionais de TI que atuam com dados e IA
• Estudantes de ciência de dados e inteligência artificial

Pre-Requisitos

Materiais

Conteúdo Programatico

Module 1: Machine Learning Fundamentals

1. ML Basics
2. ML Examples
3. Models and Examples
4. Features and Labels
5. Training, Validation Data, and Test Data
6. Training and Validation
7. Hyperparameters
8. Prediction
9. Linear Regression Models
10. Linear Regression Examples
11. Validating Linear Regression Models
12. RMSE
13. Exercise: Building a Linear Regression Model
14. Classification Models
15. Classification Examples
16. Validating Classification Models
17. Accuracy
18. Precision
19. Recall
20. Area Under Receiver Operator Curve (AUROC)
21. Exercise: Building a Classification Model
22. Neural Networks
23. Neurons, Layers, and Weights
24. Neural Network Examples
25. Exercise: Building a Neural Network

Module 2: MLOps Concepts and Pipelines

1. ML Steps
2. Data Collection
3. Feature Engineering
4. Training and Validation
5. Putting Models Into Production
6. ML Pipelines Overview
7. Batch ML Pipelines
8. Online Prediction Models
9. On-device Prediction Models
10. Streaming ML Pipelines

Module 3: Creating Data Pipelines with Dataflow

1. Dataflow Basics
2. Apache Beam Overview
3. Creating a Pipeline
4. Pipeline Basics
5. PCollections
6. Transforms
7. Input and Output
8. Dataflow Templates
9. Exercise: Programming Apache Beam Pipelines
10. Running Batch Dataflow Pipelines
11. Running Dataflow Jobs
12. Exercise: Running Dataflow Jobs
13. Streaming Dataflow Pipelines
14. Streaming Pipeline Examples
15. Exercise: Building a Streaming Pipeline with Dataflow Templates

Module 4: Feature Engineering

1. Feature Engineering Basics
2. Analyzing Features
3. Discovering Relationships in Data
4. Handling Incomplete or Invalid Data
5. One-Hot Encoding
6. Feature Engineering in Dataflow
7. Writing Transforms
8. ParDo
9. Side Inputs
10. Dataflow SQL
11. Exercise: Feature Engineering with Dataflow
12. Feature Engineering with BigQuery
13. Feature Tuning with SQL
14. Joins and Filters
15. User-Defined Functions
16. Saving Results to Tables
17. Exporting Tables
18. Exercise: Feature Engineering with BigQuery
19. Feature Engineering with TensorFlow
20. TensorFlow Transform
21. TensorFlow Schema
22. Normalizing Features
23. Bucketing Numeric Data
24. Converting Strings to Integer Vocabularies

Module 5: Training and Validating ML Models on Google Cloud

1. BigQuery ML Overview
2. Training ML Models with SQL
3. Model Types and Options
4. Data Splitting
5. Training and Validation
6. Prediction
7. Exercise: Programming Models with BigQuery ML
8. AutoML Overview
9. AutoML Tables
10. AutoML Vision
11. AutoML Natural Language
12. AutoML Video
13. Creating Datasets
14. Data Labeling
15. Training and Validation
16. Exercise: Using AutoML Vision

Module 6: Vertex AI

1. Vertex AI Basics
2. Vertex AI Workflow
3. Using Vertex AI
4. Managing Training Data
5. Datasets Overview
6. Image Datasets
7. Tabular Datasets
8. Text Datasets
9. Labeling Tasks
10. Managing Models
11. Training Models
12. Deploying Models
13. Importing Models from BigQuery ML
14. Exporting Models
15. Endpoints
16. Online Predictions
17. Batch Predictions
18. Exercise: Creating and Deploying Vertex AI Models

Module 7: Vertex AI Pipelines

1. Vertex AI Pipelines Overview
2. AI Pipeline Architecture
3. TFX vs Kubeflow Pipelines
4. Google-Provided Pipeline Components
5. Programming Custom Pipeline Components
6. Running AI Pipelines
7. Monitoring AI Pipelines
8. Exercise: Running Vertex AI Pipelines