Weekly MLOps Demo - September 3rd 2021

All Weekly Demos: #16

Recording

https://youtu.be/0V8_0naQmXM

Vision

Make GitLab the perfect companion for Machine Learning Engineers and Data Scientists.

Mission

Identify opportunities in our portfolio to explore ways where GitLab can provide a better user experience for Data Science and Machine Learning across the entire Machine Learning life cycle (model creation, testing, deployment, monitoring, and iteration).

Jupyter Notebook experience

We merged three QOL improvements for Jupyter Notebooks:

• Markers, ticks and labels in SVG images in Jupyter notebooks are not rendered
• Resize Jupyter Notebook output images to fit within the parent box
• Rendering images with relative path on Jupyter Notebooks

Working on Diff:

• Our approach will be to generate a markdown file for each ipynb version, and display as diff. We are writing the converter in Go to better integrate with our stack
  • Repository: https://gitlab.com/gitlab-org/incubation-engineering/mlops/ipynb2md
  • Issue: gitlab-org/gitlab#338890 (closed)
  • Epic: gitlab-org&6589 (closed)

First conversation with Customer

I had my first customer interview. Some highlights:

• Biggest pain point is not having Notebook Diff
• Azure DataBricks for creating and sharing notebooks, insights are moved to google docs to be shared with stakeholders
• Azure also uses MLFlow as Model Registry, which is also used by GCP.
  • We will start exploring soon #19

Customer Conversations

We are also working on creating resources to assist conversations between sales colleagues and customers on MLOps problems and solutions gitlab-com/www-gitlab-com!89355 (merged)

Up Next

• Holidays! I will be off for a week.
• Then, focus on diff gitlab-org/gitlab#338890 (closed) and MLFlow #19