AI Professional
European certification validating the operational mastery of AI: design, deployment and supervision of machine learning solutions in a professional environment.
Key skills
- ✓Design and training of ML/DL models
- ✓Deployment and MLOps in production
- ✓Ethics, bias and AI governance
- ✓Model optimisation and monitoring
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:
- →Data Scientists wishing to validate their operational skills
- →ML/AI Engineers responsible for end-to-end projects
- →AI Consultants supporting digital transformations
✗ This certification is not suitable if:
- →Beginners without Python programming experience → direct towards fundamental data science training before certification
- →Academic AI research experts seeking validation of scientific contributions → direct towards AI Fellow Series (level 901)
Prerequisites & eligibility
Open registration. Practical experience in ML recommended to optimise chances of success.
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
Advanced Modelling
Mastery of ML/DL architectures, from feature engineering to hyperparameter optimisation.
Production & MLOps
Robust deployment, CI/CD pipelines, model versioning and scalable infrastructure.
AI Governance
Bias detection, explainability (XAI), GDPR and European AI Act compliance.
Continuous Performance
Real-time monitoring, drift detection, automated retraining and A/B testing.
Exam format
Curriculum
ML/DL Fundamentals and Architectures
Supervised learning: regression, classification, decision trees, random forests, gradient boosting (XGBoost, LightGBM). Unsupervised learning: clustering (k-means, DBSCAN), dimensionality reduction (PCA, t-SNE, UMAP). Neural networks: perceptrons, backpropagation, activation functions. Deep learning architectures: CNN, RNN, LSTM, GRU, attention mechanisms, transformers. Feature engineering: selection, extraction, transformation, categorical encoding. Regularisation: L1/L2, dropout, early stopping. Validation and metrics: cross-validation, overfitting/underfitting, classification and regression metrics.
MLOps and Industrialisation
ML pipelines: orchestration with Airflow, Kubeflow, Prefect. Versioning: code (Git), data (DVC), models (MLflow, Weights & Biases). Containerisation: Docker for reproducible environments, registries. Kubernetes orchestration: scalable deployment, GPU resource management. Model serving: TensorFlow Serving, TorchServe, ONNX Runtime, FastAPI. CI/CD for ML: automated testing, model validation, progressive deployment. Infrastructure as Code: Terraform, declarative configuration. Secret and credentials management. Deployment strategies: blue-green, canary, shadow mode.
Ethics, Bias and Compliance
Types of bias: selection bias, measurement bias, sampling bias, algorithmic bias. Bias detection: demographic analysis, disparate impact, equalized odds. Mitigation: resampling, reweighting, adversarial debiasing. Explainability (XAI): SHAP values, LIME, attention visualisation, feature importance. Fairness metrics: demographic parity, equal opportunity, calibration. Model cards and documentation: transparency, limitations, use cases. GDPR: right to explanation, data minimisation, privacy by design. European AI Act: risk classification, obligations by level, prohibitions. Algorithmic audits and certifications.
Monitoring, Drift and Optimisation
Data drift: statistical detection (KS test, PSI), distribution monitoring. Concept drift: performance degradation, detection patterns. Production monitoring: business metrics, latency, throughput, errors. Observability: structured logging, distributed tracing, dashboards. Alerting: adaptive thresholds, anomaly detection. Automated retraining: triggers, pipelines, validation before deployment. A/B testing: experimental design, significance testing, multi-armed bandits. Performance optimisation: quantization, pruning, distillation, ONNX conversion. Cost optimisation: batch vs real-time, auto-scaling, spot instances. Edge deployment: TensorFlow Lite, mobile/IoT optimisation.
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 — €450 incl. VAT
Secure online payment by credit card (Stripe).
Business rate: €585 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 master all ML frameworks?
No. The exam assesses architectural concepts and cross-cutting MLOps. You must master at least one major framework (TensorFlow, PyTorch or scikit-learn), but the questions focus on principles applicable to all.
Is the ethics and bias component really assessed?
Absolutely. 20% of the exam covers AI governance, explainability and European regulatory compliance.
Does the certification cover GenAI and LLMs?
Yes. Transformer architectures, fine-tuning, advanced prompt engineering and LLM deployment (quantization, LoRA, RAG) are covered in the ML/DL architectures and MLOps module.
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.
