Intermediate / Quality Professionals 8 weeks, 5 hours per week

AI-Powered Quality Control in Production Environments

Duration 8 weeks, 5 hours per week

Level Intermediate / Quality Professionals

Seats remaining 14

Format Online lecture

14 places left — enrolment closes when full

2,750 CAD Includes defect image datasets, validation document templates, and access to a private cohort discussion channel for 6 months post-completion.

View all programs

Program outline

Program Outline

Week 1-2 — Imaging hardware, lighting, and data collection protocols
Week 3-4 — Model training and defect classification across three industry datasets
Week 5 — Threshold calibration and rejection rate analysis
Week 6 — Regulatory validation: writing protocols and audit-ready documentation
Week 7 — MES and SPC integration patterns
Week 8 — Group project: end-to-end inspection pipeline with peer review

All group projects are presented in a live session attended by participants from other cohorts, creating genuine accountability for the final output.

Ask a question

About this program

Manual visual inspection is slow, inconsistent, and expensive to scale. AI-based inspection systems can run continuously, but setting them up correctly requires more than installing a camera.

Scope of the program

This course covers the full pipeline from image acquisition through model deployment. You will learn how lighting conditions, camera placement, and image preprocessing affect model accuracy far more than algorithm choice does. The course uses real defect image libraries from metal casting, PCB assembly, and food packaging contexts.

Compliance and documentation requirements

For manufacturers operating under ISO 9001, IATF 16949, or FDA 21 CFR Part 11, AI inspection systems require specific validation documentation. A dedicated module covers how to structure validation protocols, maintain audit trails, and respond to regulator questions about algorithmic decision-making.

What you will build

By the end of the program, participants will have configured a working inspection pipeline using open-source tools, written a validation report for a simulated audit, and mapped their system against a standard quality management framework. The work is done in small groups of three to four, which mirrors how these projects actually run in production facilities.

Image acquisition setup and preprocessing techniques
Convolutional network basics for defect classification
Threshold calibration and false-positive management
Regulatory validation documentation for AI systems
Integration with MES and SPC platforms