Cloud Machine Learning
European certification validating mastery of deploying and operationalising ML models on AWS, Azure and GCP.
Key skills
- ✓Deploy ML pipelines on AWS SageMaker, Azure ML and Google Vertex AI
- ✓Orchestrate distributed training and hyperparameter optimisation in the cloud
- ✓Industrialise MLOps with CI/CD, monitoring and drift management
- ✓Optimise costs and performance of cloud ML infrastructures
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 transitioning to ML engineering
- →MLOps and ML Engineers
- →Cloud Architects specialised in AI
✗ This certification is not suitable if:
- →ML beginners without practical experience → Refer to 'Machine Learning Foundations' (Foundation level)
- →Senior ML Architects seeking research expertise validation → Consider 'AI Research & Advanced Systems' (Master 801 level)
Prerequisites & eligibility
No formal requirements but technical prerequisites strongly recommended to pass the exam.
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
Multi-cloud Deployment
Mastery of AWS, Azure and GCP managed ML services to deploy models into production.
Advanced MLOps
Orchestration of CI/CD pipelines, model versioning, performance monitoring and drift detection.
Infrastructure Optimisation
Management of compute resources, GPU/TPU, auto-scaling and cloud cost reduction strategies.
Distributed Architecture
Distributed training, parallelisation, management of massive datasets and cloud-native feature stores.
Exam format
Curriculum
Cloud ML Services and Deployment
In-depth study of managed services: AWS SageMaker (Training Jobs, Endpoints, Pipelines), Azure ML Studio (Designer, Compute Instances, Endpoints), Google Vertex AI (Training, Prediction, Pipelines). Model deployment via REST APIs, batch inference, real-time inference. Deployment strategies: blue-green, canary, A/B testing. Model version management and rollback. Integration with API Gateway, Load Balancers.
MLOps and Automation
CI/CD pipeline architecture for ML with GitLab CI, GitHub Actions, Jenkins. Kubeflow Pipelines: components, DAGs, artifacts. MLflow: tracking, registry, deployment. DVC for data versioning. Feature stores: Feast, AWS Feature Store, Vertex AI Feature Store. Orchestration with Apache Airflow, Prefect. Dependency and reproducible environment management (Docker, Conda). Metadata management and lineage tracking.
Infrastructure and Scaling
Instance selection: CPU vs GPU (NVIDIA T4, V100, A100) vs TPU. Distributed training: data parallelism, model parallelism with Horovod, PyTorch Distributed Data Parallel, TensorFlow MultiWorkerMirroredStrategy. Horizontal and vertical auto-scaling. Cost optimisation: Spot instances, Reserved instances, Savings Plans. Serverless inference with AWS Lambda, Azure Functions, Cloud Run. Quota and limit management. Storage optimisation (S3, Blob Storage, GCS).
Monitoring and Governance
Performance monitoring: latency, throughput, errors. Drift detection: data drift (KS test, PSI), concept drift, label drift. Tools: AWS Model Monitor, Azure ML Data Drift, Vertex AI Model Monitoring. Structured logging and observability with Prometheus, Grafana, CloudWatch, Application Insights. Explainability: SHAP, LIME, Integrated Gradients. GDPR compliance: right to be forgotten, pseudonymisation, audit trails. Model cards and documentation.
Optimisation and Security
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 three clouds (AWS, Azure, GCP)?
No. The exam covers multi-cloud concepts but you can pass by mastering a single cloud perfectly and knowing the equivalents on the others. About 60% of questions are agnostic, 40% specific to platforms.
What is the difference with AWS ML Specialty or Azure AI Engineer certifications?
PROVA Cloud ML is multi-cloud, European, and focuses on technological autonomy and modern MLOps practices independent of proprietary ecosystems.
Does the exam contain code?
The exam consists of 45 multiple-choice questions, in 70 minutes, online under automated supervision.
Is MLOps really assessed in depth?
Absolutely. 25% of the framework covers CI/CD, versioning, feature stores, orchestration. We assess your ability to industrialise, not just train a model. It is an ML engineering certification, not pure data science.
How to prepare effectively?
Practice on at least two of the three clouds (free tiers available), deploy models via SageMaker/Azure ML/Vertex AI, configure an MLflow or Kubeflow pipeline, and familiarise yourself with monitoring and drift detection strategies.
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.
