Agentic Quality Engineering for Embedded Systems: Quality Intelligence Across the Entire Delivery Lifecycle

Industry data shows almost 70% of embedded testing capacity is consumed by documentation, traceability matrices, and report generation, leaving only 30% for the intellectual work that creates value: risk analysis, requirements evaluation, and informed release decisions. AI has not changed this fundamental problem. It has made it faster to generate more artefacts, the faster horse problem applied to safety-critical systems.

This workshop introduces Agentic QCSD (Quality Conscious Software Delivery), an open-source framework that encodes practitioner expertise into coordinated AI agent swarms operating across the full delivery lifecycle: Ideation, Refinement, Development, Verification, and Production. Each phase has its own swarm with purpose-built agents addressing the quality questions that phase demands.

Participants work hands-on with a representative automotive ECU firmware update codebase. They run agent swarms that analyse source code, safety requirements, and test specifications directly on their laptops with no hardware required. They experience Human-in-the-Loop quality gates where agents inform but humans decide, and observe how the system builds institutional knowledge across delivery cycles.

Grounded in embedded-specific concerns including ISO 26262 traceability, timing constraints, protocol validation, and regulatory compliance, participants leave with a configured environment, full source code under MIT license, and a personal adoption roadmap.