fr
en
MLOps in practice: deploying, monitoring and maintaining an AI model
An AI model delivers lasting value only when its operation is properly managed. Use MLOps to connect deployment, monitoring and maintenance as you organise the transition to production. Strengthen your ability to coordinate data and operations teams around models monitored over time.
- Duration
- 1 day 6 hours
- Code
- IA038FR Code
Presentation
Creating a high-performing model is only the first step in an artificial intelligence project. The real challenge is making it production-ready: how do you deploy it, maintain it and sustain its performance over time? This intensive one-day course explores MLOps (Machine Learning Operations), the discipline that brings data science and IT operations together to make the model lifecycle more reliable.
The program covers the essential technical foundations for turning a prototype into a robust production solution. Learn to build automated pipelines (CI/CD), manage data versioning and orchestrate deployment through containerisation. Beyond automation, the course emphasises active monitoring to detect data drift and trigger the necessary retraining.
Through hands-on workshops using established tools such as MLflow and Docker, you will experiment with setting up a complete MLOps pipeline. You will leave with the knowledge needed to establish sound technical governance, supporting the scalability and auditability of your AI solutions.
Objectives
By the end of this course, you will be able to:
- analyse the challenges and benefits of MLOps in making AI projects production-ready;
- identify critical stages in a model's lifecycle in a production environment;
- build automation pipelines (CI/CD) for training and deployment;
- implement monitoring strategies to track performance and drift;
- apply governance, security and documentation best practices.
Program
Module 1: understanding MLOps fundamentals
- Defining MLOps and its fundamental differences from conventional DevOps.
- Analysing the specific challenges of bringing AI models into production at scale.
Hands-on exercises
- Map the complete lifecycle of an AI model within a typical organisation.
Module 2: building and automating the pipeline
- Preparing data and setting up versioning.
- Automating model training and validation stages.
- Incorporating continuous integration (CI) and continuous deployment (CD) principles.
Hands-on exercises
- Create a simplified MLOps pipeline using tools such as MLflow or DVC.
Module 3: deploying and monitoring in production
- Production deployment techniques: exposing models through APIs and containerisation with Docker.
- Monitoring performance indicators and detecting data drift.
- Lifecycle management: retraining strategies and maintenance.
Hands-on exercises
- Set up a monitoring dashboard to track model health.
Module 4: ensuring governance and scalability
- Technical documentation and auditability to ensure transparency.
- Secure access management and protection of sensitive data.
- Fundamentals of scalable cloud architectures to support workload demands.
Hands-on exercises
- Develop a structured governance plan for an MLOps project.
Audience
This course is intended for technical and operational data professionals, including:
- data scientists and data engineers seeking more professional deployment practices;
- AI project managers and technical product owners responsible for service quality;
- developers involved in integrating AI components;
- innovation managers leading the transformation of data infrastructure.
Prerequisites
The following prerequisites apply:
- Basic knowledge:
- basic machine learning knowledge is essential;
- knowledge of Python and technical project management is strongly recommended.
Teaching and assessment methods
- Initial skills assessment
- Training materials provided to participants
- Continuous assessment throughout the course
- End-of-course feedback questionnaire
- Combination of theory and practical application
- Attendance records
- Post-course follow-up evaluation
- Practical exercises
Course highlights
- Operational focus: go beyond theory to work with industry-standard tools and practices such as Docker, MLflow and CI/CD.
- Complete lifecycle: master the end-to-end process, from code versioning to production monitoring.
- Integrated governance: learn to secure your models and ensure compliance by design.
- Intensive format: develop essential practices for bringing AI into production in one focused, structured day.
Dates and sessions
Choose the date and delivery format that suit you.
No upcoming sessions are currently available.
Any brand names and logos mentioned in this course description (such as Docker, MLflow, Python and DVC) belong to their respective owners.
Their mention for educational purposes does not imply any commitment or partnership.