iSAQB CPSA-A SWARC4AI

Mô tả

Attending the iSAQB® CPSA-A Software Architecture for AI-Based Systems (SWARC4AI) course gives participants 20 Technical Competence (TC)  and 10 Methodological Competence (MC) points towards the 70 points required for eligibility for the iSAQB CPSA-A exam with Brightest. It is important to remember that as part of the 70 points required to take the iSAQB CPSA-A exam with Brightest, you will need at least ten competence points in each of the following areas: 

• Technical Competence (TC)
• Methodological Competence (MC)
• Communicative Competence (CC)

Accredited iSAQB® SWARC4AI - Software Architecture for AI-Based Systems (CPSA-A) training is based on the current iSAQB® curriculum:

Part 1 - Introduction to Software Architecture-Relevant Concepts for Artificial Intelligence

• Classify artificial intelligence, machine learning, data science, deep learning, and generative AI
• Specify typical general use cases for AI
• Know possible applications in different industries and end-user applications
• Identify risks in the application of AI
• Understanding differences from traditional software
• Decide on solutions to problems using AI or classic software development
• Know roles and their tasks as well as their cooperation in the context of AI
• Identify and prioritize AI use cases
• Know the strengths and limitations of AI
• Know the concept of the "Productivity J-Curve" in conjunction with AI technology

Part 2 -  Compliance, Security, and Alignment

• Know the influence of data protection laws on the implementation and use of AI
• Understand the objectives and regulations of the EU AI Act and their impact on the development process and architecture
• Perform classification of AI systems according to the EU AI Act risk level
• Classify copyright issues of AI-generated content
• Overview of types and degrees of openness and types of licenses of free ML models
• Understand strategies for compliance with the EU AI Act and potential challenges
• Document models and data sets for traceability and transparency
• Know security pitfalls and types of attacks on ML models
• Know and apply strategies for AI risk minimization
• Apply options for protection against attacks (AI security)
• Understanding the fundamental issues and facets of AI safety
• Understand ethical problems of AI systems and know approaches to dealing with them
• Overview of ethical guidelines
• Know the core principles of AI governance and responsible AI for companies
• Gain insight into the creation of regulatory sandboxes
• Ensuring adequate data management for the quality and security of data in AI applications

Part 3 - Design and Development of AI-Based Systems

• Understanding the life cycle of a machine learning or data science project
• Knowing process models for the software development of AI systems
• Know the types of and requirements for data and typical ML problems
• Understanding machine learning problems and their requirements
• Use input data for various AI algorithms
• Dealing with challenges such as non-determinism, data quality, and concept and model drift
• Knowing transfer learning or fine-tuning
• Selecting design patterns for AI systems
• Define the task of an ML model
• Understand and specify inputs and outputs for the functioning of an ML system
• Know metrics for measuring the performance of ML models
• Integrating ML models into existing systems
• Designing user interfaces for AI systems
• Understanding performance metrics such as latency and throughput in AI systems
• Managing scalability for increased data volumes
• Understanding robustness in AI systems and applying strategies to increase robustness
• Classify the reliability and availability of AI systems
• Understanding the reproducibility and testability of AI results
• Know security, data protection, and compliance requirements
• Classify explainability and interpretability in AI systems
• Recognizing bias in data and models
• Knowing fault tolerance in AI systems

Part 4 - Data Management and Data Processing for AI-Based Systems

• Acquiring and labeling data
• Know common platforms for publicly accessible data
• Overview of relevant tools for data labeling
• Designing efficient data pipelines and architectures
• Know strategies for data aggregation, cleansing, transformation, enrichment, and augmentation
• Overview of tools for data engineering pipelines
• Know the options for storing data

Part 5 - Important Quality Features for the Operation of AI-Based Systems

• Know (hardware) requirements for training and inference
• Know the trade-offs of different model architectures concerning quality characteristics
• Adjust different quality features of an ML model
• Understanding the costs, power consumption, and sustainable use of AI (Green IT)
• Know (hardware) requirements for training and inference
• Model training, parameters, metrics, and results tracking
• Evaluate ML models and AI systems based on them
• Know the types of drift as well as the possible causes and solutions for them
• Overview of CI/CD pipelines, model management, and deployment strategies for AI models
• Know platforms for model provision
• Classify tools for the creation of POCs of AI systems
• Knowing the deployment options of AI models
• List the advantages and disadvantages of SaaS and self-hosting
• Overview of SaaS AI solutions
• Know embedded deployments of ML models
• Understanding monitoring concerning AI-specific requirements
• Overview of sample tools for monitoring
• Understanding user feedback, methods, and tools for collecting user feedback
• Know methods for using feedback for model training
• Understanding the MLOps pipeline using a practical example (Optional)
• Make build vs. buy decisions for MLOps systems/components (Optional)
• Know MLOps tools and end-to-end platforms (Optional)

Part 6 - System Architectures and Platforms for Generative AI-Based Systems

• Overview of integration levels of AI
• Know libraries, interfaces, and tools for the integration of AI models
• Integrate AI systems into the overall architecture of an IT landscape
• Overview of relevant quality features for AI systems
• Know evaluation frameworks for AI systems (Optional)
• Discuss a case study with an imaginary professionalism (Optional)
• Fundamental understanding of generative AI
• Understanding how LLMs work
• Understand known patterns in the use of LLMs
• Knowing use cases for RAG (Retrieval-Augmented Generation)
• Knowing and understanding selected RAG techniques
• Know types of prompt engineering
• Overview of agentic workflows
• Know a selection of design patterns for generative AI systems
• Know techniques for evaluating LLM applications
• Overview of known LLMs and selection criteria
• Understanding the importance of cost management for GenAI applications
• Give examples of common libraries, interfaces, and tools related to LLM applications
• Know agentic AI software architectures, AI agent architecture components, and types of AI agent architectures (Optional)