Continuous Deployment of Machine Learning Models

Explore the implementation of continuous deployment for machine learning models in this insightful conference talk from Devoxx. Learn how Auto Trader, the UK's leading digital automotive marketplace, successfully integrated ML models into their existing continuous delivery practices. Discover the strategies and technologies used to develop a suite of new machine learning models capable of serving low-latency predictions in real-time. Gain valuable insights into the automated retraining and redeployment process using continuous deployment pipelines, leveraging technologies such as Apache Spark, Airflow, Docker, and Kubernetes. Understand the importance of a robust testing strategy to ensure model performance, accuracy, and coverage during deployments without manual intervention. This 50-minute presentation offers practical knowledge for organizations looking to streamline their machine learning model deployment process and reduce time to live for enhanced experimentation and cost-effectiveness.

Continuous deployment of machine learning models by Edward Kent & Paul Doran