AI Quality Assurance
European certification for professionals responsible for ensuring the quality, reliability and compliance of artificial intelligence systems in production.
Key skills
- ✓ML/AI model validation and testing
- ✓Bias and drift monitoring detection
- ✓Regulatory compliance (AI Act, GDPR)
- ✓AI performance metrics and observability
The question bank for this exam is prepared after your purchase. The exam is available no later than 48 hours after payment; you will receive an email as soon as it opens.
Is this right for you?
✓ This certification is right for you if:
- →QA Engineers transitioning to AI
- →Data Scientists responsible for model quality
- →MLOps and ML DevOps Engineers
- →AI governance and compliance consultants
✗ This certification is not suitable if:
- →Beginners with no experience in testing or data: first consider fundamental training in QA or Python for data
- →AI Architects seeking a high-level strategic certification: prefer the Master 801 or Fellow Series certifications in AI architecture
Prerequisites & eligibility
Minimum 18 months professional experience in testing, data science or MLOps recommended.
Purchase options
Exam voucher, LMS pathway, bundle packs — build your order
Options d'achat
Composez votre parcours — voucher d'examen, préparation LMS, packs entreprise.
Skills assessed
Systematic AI Testing
Mastery of testing frameworks for ML models and AI pipelines in production.
Bias Detection
Identification and mitigation of algorithmic biases according to European standards.
Continuous Monitoring
Monitoring of model degradation and alerts on data drift.
AI Act Compliance
Application of European regulatory requirements for critical AI systems.
Exam format
Curriculum
Fundamentals of AI Testing
Validation strategies for supervised and unsupervised models. Splitting techniques (train/test/validation). K-fold cross-validation. Performance metrics by problem type (classification, regression, clustering). Data quality assessment and anomaly detection in datasets. Preprocessing pipeline testing. Robustness and generalisation evaluation. Non-regression tests for retrained models.
Bias Detection and Mitigation
Bias taxonomy: selection bias, measurement bias, aggregation bias, representation bias. Fairness metrics: demographic parity, equalized odds, disparate impact. Debiasing techniques: pre-processing, in-processing, post-processing. Audit tools: AI Fairness 360, Fairlearn. Documentation of known limitations and biases. Compliance with European ethical guidelines. Independent algorithmic audit process.
Monitoring and Observability
Data drift detection (Kolmogorov-Smirnov, Population Stability Index). Concept drift and model decay. Monitoring architecture: logs, metrics, alerts. Model health dashboards. Shadow mode and canary deployments. A/B testing for ML models. Tools: MLflow, Weights & Biases, Evidently AI, Grafana for ML. Automatic rollback strategies. Prediction incident management.
Compliance and Governance
European AI Act: system classification (unacceptable risk, high, limited, minimal). Requirements for high-risk systems: data, documentation, transparency, human oversight. GDPR and automated decisions (article 22). Explainability and right to explanation. Model registers and traceability. Mandatory technical documentation. Compliance assessment process. Role of notified bodies.
QA Tools and Automation
Great Expectations framework for data validation. DeepChecks for ML model tests. Pytest and pytest-ml for automated testing. CI/CD for ML with GitHub Actions, GitLab CI. Validation gates in MLOps pipelines. Infrastructure as code for test environments. Dataset and model version management. Load and performance tests for ML endpoints.
How to fund your certification
3 ways to cover the cost of your certification
OPCO — Employer funding
Ask your employer to cover the cost through your OPCO (employee training plan).
Personal — €420 incl. VAT
Secure online payment by credit card (Stripe).
Business rate: €550 excl. VAT
Business — custom quote
Teams, cohorts or private catalogue: a quote tailored to your volumes and goals.
Frequently asked questions
Do I need to be a developer to take this certification?
No. A QA profile with basic Python knowledge is sufficient. The exam assesses understanding of testing methods, anomaly detection and compliance, not the development of complex models.
Which tools are covered in the exam?
The exam covers transversal concepts applicable to any ecosystem: Great Expectations, Evidently AI, MLflow, Weights & Biases, DeepChecks, as well as CI/CD best practices for ML.
Is this certification recognised for MLOps positions?
Yes, it perfectly complements an MLOps profile by adding the quality and compliance dimension. It is particularly valued in regulated environments (finance, health, public sector) where AI governance is critical.
Your certification pathway
Before, during, after — the recommended progression
Droits du candidat
Transparence
Les critères de décision et le barème sont documentés et accessibles avant l'examen.
Droit d'appel
Toute décision peut être contestée dans les 30 jours. Examen par un tiers indépendant.
Plainte
Toute personne peut signaler un dysfonctionnement. Formulaire public accessible sans compte.
