With MLOps, you optimize your model development and deployment of your data.

A practical introduction and comprehensive insight into MLOps.

This two-day training provides data science specialists with a hands-on introduction to MLOps, optimizing the entire machine learning lifecycle through automation and standardization. With a focus on code and data versioning, as well as model parameter monitoring, it ensures consistent and reproducible results. Continuous Integration and Continuous Deployment (CI/CD) accelerate the launch of new ML solutions and enhance the efficiency of the development process.

Our training team consists of professionals who are both active in machine learning research and have extensive experience in the practical implementation of data models in companies.

Outcomes

Refresh the basics of machine learning and terminology
Learn techniques for data acquisition, preparation, and versioning
Understand and develop MLOps processes: model training and development using pipelines
Use DVC for data versioning and CML for ML pipelines
Get to know other tools like MLFlow and Kubeflow in the MLOps ecosystem

Prerequisites

Participants need a laptop with direct internet access

Target audience

Our training is aimed at people who work with data and data models and already have prior knowledge in data science, and who want to learn techniques and processes to deploy and maintain services from these data and models.

Conditions

CHF 1900 / participant for two days.
Group sessions consisting of 8 to 24 participants.
Includes catering and documentation.
Discounts available for groups of 12 participants or more.

Content

Our trainings consist of diverse presentations and hands-on labs to deliver the content in an engaging way. We are happy to refer to your infrastructure in consultation. If additional content is needed, we can make adjustments upon your request.

MLOps Data Science

Recap of machine learning, model types, and their application areas
Data collection and preparation using various techniques
Model training and tuning with Code Spaces/GitHub Actions in the Free Tier
Problem statement: making the process reproducible and measurable
From prototype to pipeline
ML pipelines and testing
Data versioning
Metrics and experiments

Companies

Individual company trainings are possible. Contact us for costs and dates.

Meet two of your trainers

Sigve Haug

Trainer

Sigve is the director of studies at the Mathematical Institute (MAI) of the University of Bern.

Iwan Imsand

Trainer

Iwan likes the quote from Don Draper: Make it simple, but significant.

Here’s what our clients say about our trainings:
Nathanael Weber, Bern

«The training was exceptionally well tailored to our needs. The practical exercises were just difficult enough that more questions about Helm arose and were answered competently by the instructor. Many thanks acend for this experience!»