What this book covers
Chapter 1, Navigating the ML Lifecycle with ML Solutions Architecture, introduces ML solutions architecture functions, covering its fundamentals and scope.
Chapter 2, Exploring ML Business Use Cases, talks about real-world applications of AI/ML across various industries such as financial services, healthcare, media entertainment, automotive, manufacturing, and retail.
Chapter 3, Exploring ML Algorithms, introduces common ML and deep learning algorithms for classification, regression, clustering, time series, recommendations, computer vision, natural language processing, and generative AI tasks. You will get hands-on experience of setting up a Jupyter server and building ML models on your local machine.
Chapter 4, Data Management for ML, addresses the crucial topic of data management for ML, detailing how to leverage an array of AWS services to construct robust data management architectures. You will develop hands-on skills with AWS services for building data management pipelines for ML.
Chapter 5, Exploring Open-Source ML Libraries, covers the core features of scikit-learn, Spark ML, PyTorch and TensorFlow, and how to use these ML libraries for data preparation, model training, and model serving. You will practice building deep learning models using TensorFlow and PyTorch.
Chapter 6, Kubernetes Container Orchestration Infrastructure Management, introduces containers, Kubernetes concepts, Kubernetes networking, and Kubernetes security. Kubernetes is a core open-source infrastructure for building open-source ML solutions. You will also practice setting up the Kubernetes platform on AWS EKS and deploying an ML workload in Kubernetes.
Chapter 7, Open-Source ML Platforms, talks about the core concepts and the technical details of various open-source ML platform technologies, such as Kubeflow, MLflow, AirFlow, and Seldon Core. The chapter also covers how to use these technologies to build a data science environment and ML automation pipeline.
Chapter 8, Building a Data Science Environment Using AWS ML Services, introduces various AWS managed services for building data science environments, including Amazon SageMaker, Amazon ECR, and Amazon CodeCommit. You will also get hands-on experience with these services to configure a data science environment for experimentation and model training.
Chapter 9, Designing an Enterprise ML Architecture with AWS ML Services, talks about the core requirements for an enterprise ML platform, discusses the architecture patterns and best practices for building an enterprise ML platform on AWS, and dives deep into the various core ML capabilities of SageMaker and other AWS services.
Chapter 10, Advanced ML Engineering, provides insights into advanced ML engineering aspects such as distributed model training and low-latency model serving, crucial for meeting the demands of large-scale model training and high-performance serving requirements. You will also get hands on with distributed data parallel model training using a SageMaker training cluster.
Chapter 11, Building ML Solutions with AWS AI Services, will introduce AWS AI services and the types of problems these services can help solve without building an ML model from scratch. You will learn about the core capabilities of some key AI services and where they can be leveraged for building ML-powered business applications.
Chapter 12, AI Risk Management, explores AI risk management principles, frameworks, and risk and mitigation, providing comprehensive coverage of AI risk scenarios, guiding principles, frameworks, and risk mitigation considerations across the entire ML lifecycle. It elucidates how ML platforms can facilitate governance through documentation, model inventory maintenance, and monitoring processes.
Chapter 13, Bias, Explainability, Privacy, and Adversarial Attacks, delves into the technical aspects of various risks, providing in-depth explanations of bias detection techniques, model explainability methods, privacy preservation approaches, as well as adversarial attack scenarios and corresponding mitigation strategies.
Chapter 14, Charting the Course of Your ML Journey, outlines the stages of adoption and presents a corresponding maturity model designed to facilitate progress along the ML journey. Additionally, it addresses key considerations essential for overcoming the hurdles encountered throughout this process.
Chapter 15, Navigating the Generative AI Project Lifecycle, discusses the advancement and economic impact of generative AI, the various industry trends in generative AI adoption, and guides readers through the various stages of a generative AI project, from ideation to deployment, exploring various generative AI technologies, and limitations and challenges along the way.
Chapter 16, Designing Generative AI Platforms and Solutions, explores generative AI platforms’ architecture, the retrieval-augmented generation (RAG) application architecture and best practices, considerations for generative AI production deployment and practical generative AI-powered business applications across diverse industry use cases.
The chapter finishes with a discussion on artificial general intelligence (AGI) and various theoretical approaches the research community has taken in their pursuit of AGI.
