The Full Machine Learning Lifecycle -How to use Machine Learning in Production (MLOps)
Organized by D ONE Solutions AG
This workshop will provide an introduction to MLOps. After attending the workshop, the participants will have gained an understanding of MLOps principles and will have applied these hands-on to an end-to-end data science project using real-world data and state-of-the-art frameworks.
MLOps is a Machine Learning (ML) engineering practice that aims to unify ML system development (Dev) and ML system operation (Ops). It serves as a counterpart to the DevOps practice in classical software development which involves Continuous Integration (CI) and Continuous Deployment (CD). Practicing MLOps advocates automation and monitoring at all steps of ML system construction, including integration, testing, releasing, deployment and infrastructure management.
In this workshop, we discuss how the DevOps principles from software engineering translate to data science and machine learning by introducing all steps of an MLOps workflow.
In the practical part of the workshop we will discuss each of these steps in detail and use a real-world windpark dataset to show how each stage may be implemented using state-of-the-art open-source frameworks. We will demonstrate how different frameworks like Data Version Control (DVC), MLflow and Apache Airflow complement each other and how they may be combined for an end-to-end solution.
This workshop targets professionals working with data. For the theoretical part no prior knowledge of MLOps is required as we will explain all necessary concepts. Following the practical part requires knowledge of Python and basic knowledge of the Machine Learning domain. The frameworks used in this workshop will be introduced to the audience during the workshop and no prior experience is necessary to benefit from the session. All participants need to bring their own laptops. Participants will need to have admin rights to their laptops so they can install VS Code (or any other tool to connect via SSH to a Virtual Machine).
After the workshop, all participants are invited to join us at a networking lunch (12:30 – 13:30) together with the participants of the Full Machine Learning Lifecycle workshop.
MLOps lifecycle (source)
Time frame: June 22, 8:30 – 12:30
1. Why MLOps is needed
- ML lifecycle
- MLOps Theory (CI, CT, CD)
- Commercial solutions from big data providers
2. Practical Hands-on
- Pipeline orchestration
- Data versioning and preprocessing
- Continuous model training, tracking and serving
- Deploying in production