Data Version Control and uv: Git reproducible data science from POC to prod

2 tech talks on data science tools.

All experience levels welcome.

Talk 1: DVC - Data Version Control has been central to enabling the Continuum Media Data Science team to rapidly prototype trained machine learning models while still being entirely reproducible on any machine by any team member. We've successfully extended our git version control processes with our experiment process in a production ready pipeline that deploys AWS SageMaker endpoints ready for inference via our online platform. In this talk we'll introduce DVC, share some example code and outline how you can use it to achieve the same in your work.

Break: pizza sponsored by Continuum Media

Talk 2: uv is a package and project manager for python that takes a radically different approach. From astral.sh (the same people who made ruff!), uv solves package environments 10-100 times faster than other tools, and also has the potential to replace multiple tools in your developer environment into a single unified system. This talk will introduce uv, and enable you to get started using it in your projects.

Continuum Media have graciously sponsored this event. Continuum Media builds the only managed service exchange for linear TV, and we build it right from Cardiff.