DevOps Paradox

DOP 299: Managing Your AI Workloads With KitOps

Jan 22, 2025

Gorkem Ercan, CTO at Jozu, sheds light on KitOps, an open-source initiative revolutionizing AI/ML model management. He discusses how ModelKits enhance efficiency, surpassing traditional Docker containers. Gorkem tackles challenges like standardization and compliance with regulations such as GDPR. The conversation highlights the importance of experimentation tracking and how KitOps simplifies workflows, making deployment more intuitive for data scientists. Tune in for insights on bridging the gap between tech creators and users!

42:00

Creator website

Episode guests

Gorkem Ercan

AI Summary

AI Chapters

Episode notes

Podcast summary created with Snipd AI

Quick takeaways

KitOps standardizes the management of AI and ML models by using OCI artifacts, enhancing packaging, deployment, and tracking across environments.

The innovative model kit approach facilitates targeted access to datasets and models, reducing unnecessary data transfer and improving operational efficiency for teams.

Deep dives

The Evolution of AI Workflows with KitOps

KitOps introduces a structured approach to manage AI and machine learning project artifacts by utilizing OCI (Open Container Initiative) artifacts. This method allows data scientists to package their datasets, models, and related code into what are termed 'model kits,' which can be stored in OCI registries like Docker Hub. The primary advantage of this system lies in its immutability, ensuring that different versions of data and models are preserved, making tracking and management significantly easier. Additionally, by leveraging container tooling, KitOps facilitates integration into existing DevOps pipelines, enhancing both standardization and productivity across teams.

Intro

2min

Introduction to KitOps for AI Workloads

24min

Enhancing Kit CLI with Plugin Capabilities for AI Management

2min

The Essentials of Experimentation Tracking in AI/ML

7min

Bridging the Gap Between Model Creators and Users

2min

Streamlining AI Workflows with Jozu and KitOps

5min

#299: The ability to efficiently manage and analyze data is crucial in today's rapidly evolving tech landscape. One innovative solution that addresses this need is ModelKit. ModelKits are built on existing standards, ensuring compatibility with the tools your data scientists and developers already use.

In this episode, Darin and Viktor speak with Gorkem Ercan, CTO at Jozu, about KitOps, the open source DevOps project built to standardize packaging, reproduction, deployment, and tracking of AI / ML models, so it can be run anywhere, just like application code.

Gorkem's contact information:

X (Formerly Twitter): https://x.com/GorkemErcan

LinkedIn: https://www.linkedin.com/in/gorkemercan/

YouTube channel:

https://youtube.com/devopsparadox

Review the podcast on Apple Podcasts:

https://www.devopsparadox.com/review-podcast/

Slack:

https://www.devopsparadox.com/slack/

Connect with us at:

https://www.devopsparadox.com/contact/