top of page
iStock-1198781366.jpg

Building an end to end deployment process for a SageMaker Machine Learning model

User Journey Mapping and Business Improvement

How might we design the day one end to end process for SageMaker so that all our users can work together to deliver a successful deployment

Overview

The Challenge...​

Continuing from the previous research Paths to Production, this is a real use case of creating an end to end deployment process for a SageMaker Machine Learning(ML) Model to be deployed by cross functional teams.

Our approach

  • A define phase to...

    • find the alignment between past research and current

    • understand the use case

  • A workshop phase to...

    • Co-design what the target state process might look like 

My Role

My role in this engagement was to conduct design research, conduct interviews, faciliate workshops and map out the journey. 

Define

Define

Utilising past research in the present:

​There were essentially three artefacts that we could use for this piece; the research from Paths to Production, existing data science process maps and a draft of the future state journey. This was extremely useful and formed as the foundation for all our future work. 

The first step was to translate the journey into miro and use it to run our first workshops with the wider working group. This working group consisted of the Head of Data Science, Head of DevSecOps, Data Scientists, Machine Learning Engineers, Solution Architects and Software Engineers. The below journey was what was captured during our first 2 sessions.

Workshop

Workshop

Through November to January, we conducted a total of: 

  • 10 Co Design Workshops 

  • 8 interviews with individual stakeholders

  • 3 knowledge sessions to understand the scope and use case

The structure of these workshops built upon the previous, we would

  • run through any updates since we last spoke

  • areas where clarification is needed

  • build out new areas 

  • validate with the wider team

Target State Journey

Target State Journey

The intention of this piece (which we succeeded) was to map out the end to end deployment process for a SageMaker Machine Learning(ML) Model to be deployed by cross functional teams.

If you would like to see the journey in detail, please click on the below diagram for the pdf.

bottom of page