176: MLOps at SwampUp
from Programming Throwdown Hosted by Patrick Wheeler and Jason Gauci
James Morse: Software Engineer at Cisco System Administrator to DevOps Difference between DevOps and MLOps Getting Started with DevOps Luke Marsden: CEO of Helix ML How to start a business at 15 years old BTRFS vs ZFS MLOps: the intersection of software, DevOps and AI Fine-tuning AI on the Cloud Some advice for folks interested in ML Ops Yuval Fernbach: CTO MLOps & JFrog Starting Qwak Going from a jupyter notebook to production ML Supply Chain Getting started in Machine Learning Stephen Chin: VP of DevRel at Neo4J Developer Relations: The Job What is a Large Language Model? Knowledge graphs and the Linkage Model How to Use Graph databases in Enterprise How to get into ML Ops.
Embed this episode
Chapters
- 0:15 Introduction to the episode and MLOps
- 8:17 Introduction
- 17:34 Learning Programming Basics and Tools
- 22:27 The Evolution of Automated Systems
- 27:46 Challenges of data versioning in ML
- 31:16 ZFS vs. Btrfs file system mechanics
- 35:09 ZFS vs. Btrfs
- 42:15 The Rise of Local LLMs and Fine-Tuning
- 46:08 Market Validation and Fine-Tuning Success
- 51:54 The mechanics of API calling in LLM systems
- 55:54 Introduction to Yuval Fernbach
- 59:51 Yuval's background and experience at AWS
- 1:03:49 The Need for Platform Engineering
- 1:07:45 Data Consistency in ML Systems
- 1:10:44 Challenges of Data Pipelines
- 1:14:40 Transitioning from Notebooks to Production
- 1:24:39 Advice for Students in ML
- 1:29:12 Defining Developer Relations
- 1:35:39 Hacking the LLM
- 1:41:47 LLMs in the real world
- 1:45:55 The limitations of standard RAG vs. Knowledge Graphs
- 1:52:35 Outro
No transcript yet.