Add new blog post: How LY Corporation Utilizes MLflow

[yukimori]

This Pull Request adds a new blog post titled “How LY Corporation Utilizes MLflow”.

The article introduces how LY Corporation provides a managed MLflow service on our internal AI platform,
including:

• Our architecture for running MLflow instances on Kubernetes
• Service-to-service authentication and authorization using OAuth 2.0 with Athenz
• Usage insights and trends based on internal MLflow activity

The goal of this post is to share our experience integrating MLflow at scale within a large organization,
and to offer practical insights for other teams operating MLflow in production environments.

Please let me know if you would like any clarification or adjustments.

Although this post has already received an initial review through Shiga-san, I have made the following updates (apologies for the changes after the review):

• Based on advice from the Athenz team, I revised the explanation of the service-to-service authentication and authorization mechanism to use terminology aligned with the standard OAuth 2.0 specification.
• Updated the graph to show the average based on calendar days instead of business days.

[github-actions]

🚀 Netlify Preview Deployed!

Preview URL: https://pr-408--test-mlflow-website.netlify.app

Details

PR: #408
Build Action: https://github.com/mlflow/mlflow-website/actions/runs/19664149326
Deploy Action: https://github.com/mlflow/mlflow-website/actions/runs/19850245858

This preview will be updated automatically on new commits.

[B-Step62]

Thank you so much for the great blog with tons of learnings, @yukimori-san! Left comments but mostly minor formatting and styles.

[B-Step62]

We can skip the self-introduction since the author name and position will be shown at the top of the page. (but it's totally fine to keep the company link).

[yukimori]

I removed the self-introduction as follows.

I’m responsible for building and operating the Managed MLflow service for the internal AI platform at LY Corporation.

[B-Step62]

@yukimori I wonder we can show this number first before auth sections to let readers get the scale of the platform.

[yukimori]

I added numerical information at the end of the section “How LY Corporation Uses MLflow: An Overview.”

This platform provides end-to-end MLOPs capabilities from model development to deployment and operation. MLflow is a core component of the platform for managing model training, evaluation, and model lifecycle.
Currently, MLflow is used by approximately 40 services, and it handles more than 600,000 access requests per day on average.

[yukimori]

@B-Step62

Thank you for your comment.

I have merged your suggestions and made revisions in accordance with the comments. I would appreciate it if you could review it again.

[yukimori]

@B-Step62

Just a gentle reminder on this PR.
When you have a moment, I’d really appreciate your review.