Fundamentals of Artificial Intelligence

Overview

The AI Quality Engineering (AI-QE) course delivers a system-level understanding of how modern AI systems—particularly Large Language Models are architected, trained and executed in production, from a Quality Engineering perspective.
Instead of treating AI as a black box, the course breaks it down into its core computational primitives, including tokens, embeddings, parameters, transformer layers, attention mechanisms, context windows, training data, loss functions, optimizers, inference pipelines, decoding strategies and evaluation metrics. Students will analyze how these components interact, how variability is introduced at each stage and how defects propagate across the end-to-end AI lifecycle.
The course focuses on testing AI as a probabilistic system, where outputs are non-deterministic and quality must be assessed across distributions rather than fixed expectations. Participants will learn how to design robust validation strategies, evaluate outputs for correctness, consistency, bias and hallucination and apply both classical and AI-specific metrics to measure system performance.
By bridging AI fundamentals with real-world QA practices, this course equips engineers to systematically test, evaluate and reason about AI-driven applications, including LLM-based systems such as chatbots, copilots and generative platforms.
Through real-world QA examples and guided exercises, participants will explore how AI is transforming software testing - from intelligent test generation and defect prediction to model validation and automation strategies
This course also provides a structured roadmap to help you become interview-ready for AI QA roles, combining theoretical foundations with real-world testing perspectives.

Students will gain working knowledge of

• Model - structure and behavior of AI systems
• Token - how text is segmented and processed
• Embeddings - vector representations enabling semantic understanding
• Parameters - learned weights encoding model knowledge
• Transformer Layers (Blocks) - stacked architecture enabling deep learning
• Attention Mechanism - how models focus on relevant context
• Context Window - limits of input memory and its impact on reasoning
• Training Data - data quality, bias, and coverage considerations
• Loss Function - how models measure error during training
• Optimizer - how models update parameters (e.g., gradient descent variants)
• Inference - how models generate outputs in real time
• Fine-Tuning & Alignment - adapting models to tasks and human preferences
• Decoding Strategies - controlling randomness and output quality
• Prompt - structured input as a control interface
• Output Tokens - generated responses and their constraints
• Evaluation Metrics - quantitative and qualitative validation methods

After this class, you will be able to

System Understanding

• Differentiate between key types of AI - including Narrow AI, General AI and Superintelligent AI
• Understand Narrow AI subcategories, such as Generative AI, Reactive Machines, Limited Memory AI, Natural Language Processing (NLP), Computer Vision, Speech Recognition, and Robotics
• Describe how Generative AI works in text, image, audio, video, and code generation applications
• Decompose an AI system into its functional components and explain how they interact
• Identify failure points across the AI pipeline (Data → Training → Inference → Output)
• Explain how architectural elements like attention and context window impact model behavior
• Identify and compare modern AI models, including Large Language Models
• Analyze real-world applications of AI and discuss trends shaping its future
• Understand the architecture and core components of AI systems, including models, parameters, tokens and datasets
• Recognize how AI technologies are integrated into modern QA tools, improving automation and test generation
• Evaluate and validate AI-based software systems through QA-focused test scenarios and hands-on analysis
• Build confidence to discuss AI testing concepts and practices in technical interviews and professional discussions
• Machine learning frameworks (PyTorch)