AI Assisted Software Delivery Training Course

Module 1: Context, Scope, and Delivery Challenges

• Autocomplete capabilities versus autonomous multi-step execution
• Common misconceptions regarding AI in software delivery
• Why enhanced prompting alone is insufficient
• Assessing participant tooling, pain points, and objectives
• Selecting the appropriate AI operating model for engineering teams

Module 2: Specification Ingestion and Structured Decomposition

• Creating a structural inventory of stakeholder documentation
• Techniques for extracting requirements
• Chunking strategies: structural, semantic, and sliding-window approaches
• Maintaining dependencies and cross-references
• Handling tables, diagrams, flowcharts, and mixed-media inputs
• Efficient management of context windows

Module 3: Human Judgment Boundaries

• Identifying where human decision-making remains essential
• Recognising hallucinated dependencies
• Detecting fabricated constraints and inverted logic
• Avoiding unsafe default behaviours
• Validation frameworks for traceability, consistency, and completeness

Module 4: From Requirements to Code with Agentic Tools

• Implementing an architecture-first delivery model
• Component mapping and defining service boundaries
• Utilising API contracts as delivery anchors
• Establishing persistent rules and constraints in AI tools
• Linking task instructions directly to requirements
• Minimal prompting versus constrained prompting strategies
• Contract-first generation for backend and frontend components

Module 5: Agentic Iteration Loop

• The self-correction spiral
• Managing controlled iterative delivery cycles
• Reviewing diffs and code changes
• Identifying scope creep and unauthorised modifications
• Managing limited context memory
• Leveraging iteration history for continuous improvement

Module 6: Code Quality Enforcement

• Applying prompt constraints to edge cases
• Using rules documents as living governance artefacts
• Implementing automated gates with linting and static analysis
• Security scanning of AI-generated code
• Checking dependency and architecture conformance
• Human review protocols for AI outputs

Module 7: Feedback Loops and Continuous Improvement

• Integrating structured failures back into AI workflows
• Defining bounded iterations and stop criteria
• Logging cycles and outcomes
• Refining rules documents over time
• Building reusable engineering intelligence

Module 8: Security Anti-Patterns in AI Delivery

• Common security risks in generated code
• Technology-specific security rule appendices
• Pre-commit security scanning
• Secure SDLC controls for AI-assisted development
• Maintaining human accountability in secure delivery

Module 9: Testing Anchored to Specifications

• Generating test specifications from requirements
• Designing tests in domain language
• Safely generating test implementations
• Concepts of mutation testing
• Validating specification coverage
• Reviewing assertion strength
• Diagnostic questioning models

Module 10: Maintaining the System

• Managing living artefacts: contracts, maps, rules, and test specs
• Evolving constraints over time
• AI governance for long-term maintainability
• Preventing technical debt using AI controls
• Operating models for sustainable AI engineering teams