ITTA offers Train and deploy a machine learning model with Azure Machine Learning (DP-3007), an official Microsoft Learn path. The course covers Azure ML workspace creation, ML model training and deployment pipelines, AutoML and no-code designer usage, MLflow experiment tracking, and model deployment to online and batch endpoints. It targets data scientists, data engineers, MLOps engineers and cloud architects industrialising machine learning projects on Azure. Delivered in Geneva, Lausanne and interactive virtual classroom.
FILTER BY
- Domains
- Editors
- Location
- Format
- Level
- Certifying
- Confirmed training
Build a data warehouse in Microsoft Fabric with this 1-day training in Geneva and Lausanne. Learn T-SQL modelling, star schemas, and cross-database queries at ITTA with MCT trainers.
Certifying
Companies starting their machine learning journey generally go through three phases. Phase one: an isolated data scientist trains models in a local Jupyter notebook, saves files on a network share, and no one really knows which model is in production. Phase two: the team discovers Git, MLflow and Docker containers, then assembles a patchwork platform with a few bash scripts and a lot of stress at update time. Phase three, MLOps maturity: the organisation chooses a managed platform covering the entire cycle (data versioning, training, model registry, deployment, monitoring, governance) with proven components and centralised access and cost management. Azure Machine Learning is one of the three major managed platforms in this category, alongside Amazon SageMaker and Google Vertex AI.
In French-speaking Switzerland, MLOps emerges strongly in organisations that have validated their first AI use cases (banking, insurance, university health, watchmaking industry, logistics) and need to move from isolated POC to controlled production. Azure Machine Learning is particularly adopted by organisations already committed to the Microsoft Azure ecosystem and seeking consistency with their other cloud services (Azure DevOps, Azure Data Lake, Azure Synapse).
Our Azure Machine Learning training in the ITTA catalogue:
DP-3007 is a short and operational official Microsoft Learn path. The course addresses Azure ML workspace creation, resource organisation (compute, datastores, environments), data import and versioning, model training via several approaches (Python script, drag-and-drop no-code designer, AutoML), experiment tracking with integrated MLflow, model registration in the model registry, and deployment to online endpoints (real-time predictions) and batch endpoints (massive processing).
Azure Machine Learning is part of a coherent set of Microsoft Azure AI certifications. The AI-102 certification (Designing and Implementing an Azure AI Solution) covers Azure AI Services usage (Computer Vision, Document Intelligence, Speech, Language) and includes an Azure Machine Learning dimension for custom model deployment. The DP-100 certification (Designing and Implementing a Data Science Solution on Azure) is a widely used tool for Azure data scientists and covers Azure Machine Learning in depth, from data preparation to deployment and monitoring.
DP-3007 is positioned as an operational introduction to Azure Machine Learning, complementary to AI-102 for data science profiles and preparatory to DP-100 for those targeting the full certification.
One specificity of Azure Machine Learning is offering three very different approaches to train a model, suited to varied profiles.
AutoML automatically generates candidate models from a dataset and a target task (classification, regression, forecasting, computer vision, NLP). The system tests several algorithms and configurations, optimises hyperparameters and proposes the best models with their evaluation metrics. This is the recommended approach to start quickly or for business analysts without deep data science background.
No-code designer is a drag-and-drop interface allowing you to visually assemble an ML pipeline (preparation, training, evaluation) without writing code. Suited for analytical profiles wanting to understand the ML mechanics without mandatory Python.
Python SDK (Azure ML v2) is the classic data scientist approach: Python code with the azure-ai-ml libraries, custom training, full control. This is the path used in production for industrialised pipelines.
The course addresses all three approaches and helps identify which is relevant for your context.
Our Azure Machine Learning audience is diverse: data scientists already comfortable with Python and scikit-learn discovering MLOps industrialisation on Azure, data engineers building reproducible training pipelines, MLOps engineers in charge of deployments, monitoring and cloud costs, Azure cloud architects integrating an ML dimension into their landscape, AI project managers scoping an ML project with their technical team, and business analysts wanting to understand concretely what ML can bring to their domain through AutoML or the designer.
Azure Machine Learning fits into the broader AI and data ecosystem. The data science and applied AI sub-domain regroups our data science and operational AI training. For profiles covering all Azure AI services, Azure AI Services covers Computer Vision, Document Intelligence, Speech, Language and Azure OpenAI Service. For the Python language used in Azure ML, Python offers our Python training oriented to data and ML.
On the editor side, the stack is anchored to Microsoft Azure, regrouping our training on the entire cloud platform. For organisations combining ML, BI and data engineering, Microsoft Fabric brings a unified platform articulating with Azure ML for deployment.
ITTA offers a training catalogue with Modern Work, Data & AI and Security specialisations. Our Azure trainers are certified trainers active on Azure projects in French-speaking Switzerland, ensuring consistency with official Microsoft Learn materials and fresh content against the platform’s very fast evolution. Open-enrolment sessions run in Geneva and Lausanne, and in interactive virtual classroom. For data teams seeking in-house training tailored to their Azure tenant and use cases, we adapt exercises to your real context.
Do I need to master Python before DP-3007?
A Python base is recommended but not mandatory. The course covers AutoML and the no-code designer, which do not require Python. To fully leverage the Python SDK and industrialise pipelines, a solid Python base (NumPy, pandas, scikit-learn) is strongly recommended.
Azure ML vs Databricks vs Synapse: which to choose?
The three Microsoft services (Azure ML, Azure Databricks, Azure Synapse) overlap but have distinct positioning. Azure ML is the MLOps platform dedicated to the full ML cycle. Databricks excels on big data with Spark and also offers an ML environment. Synapse is the analytics and data warehouse platform. The course addresses these differences to help you articulate these services.
What costs should I expect on Azure ML?
Costs depend mainly on compute clusters used for training and deployed endpoints. A central point is automatic shutdown management for unused compute, addressed in class. Azure ML workspaces themselves are free.
Does Azure ML support large language models?
Yes, Azure ML integrates with Azure AI Foundry (ex-Azure AI Studio) for fine-tuning and deploying large language models. For pure LLM use cases (chatbots, RAG, generation), Azure AI Foundry is generally the preferred path, and Azure ML remains central for classic machine learning and custom models.
Our Azure ML sessions are scheduled in Geneva and Lausanne, and in interactive virtual classroom with a live trainer. Group sizes stay small, allowing the trainer to tailor examples to your sector context (banking, insurance, healthcare, industry, public sector). An active Azure subscription (or trial account) is required for hands-on exercises, and our trainers brief you in advance. For data teams training several collaborators as a cohort, we offer in-house sessions calibrated to your real Azure landscape.
Monday to Friday
8:30 AM to 6:00 PM
Tel. 058 307 73 00
ITTA
Route des jeunes 35
1227 Carouge, Suisse
Monday to Friday, from 8:30 am to 06:00 pm.
