Per Kops: Architecture of Ollama and Local LLMs - Episode 342

Azure & DevOps Podcast

Per is an experienced software architect and developer with more than 19 years of experience in designing, developing, and testing software. He is currently serving as a Distinguished software developer at Delegate, a Microsoft Partner in Denmark.

Topics of Discussion:

[3:02] More about Per’s background in nearly two decades in .NET and Azure consulting.

[6:11] The concept of a boutique consultant.

[7:00] Founding ATC-Net to share common .NET libraries and avoid duplicative coding.

[10:02] Building trust with customers by open-sourcing “generic” components.

[11:50] Creating a connector to run Olama (local LLMs) with Semantic Kernel.

[20:08] Why Per “fell in love” with Semantic Kernel’s abstraction layers and extension possibilities.

[21:44] The role of vector stores (like Azure Cognitive Search) in retrieval-augmented generation.

[24:33] What do we mean by “vector” in the AI space?

[26:39] What is a “text chunker”?

[31:43] Azure AI search.

[34:07] Finding the right tool for the right task.

Mentioned in this Episode:

Clear Measure Way

Architect Forum

Software Engineer Forum

Programming with Palermo — New Video Podcast! Email us at [e-mail].

Clear Measure, Inc. (Sponsor)

.NET DevOps for Azure: A Developer’s Guide to DevOps Architecture the Right Way, by Jeffrey Palermo

atc-net GitHub

Per Kops LinkedIn

Per Kops GitHub

Want to Learn More?

Visit AzureDevOps.Show for show notes and additional episodes.