azure-openai-llm-notes: A Curated, Monthly-Updated Resource List for Azure OpenAI and LLMs

What This Repository Covers and Who Needs It

Azure OpenAI and large language model engineering have produced a large volume of tooling, research papers, tutorials, and platform documentation. Finding relevant material requires either following multiple sources or maintaining a personal reading list. azure-openai-llm-notes takes on that curation work for a community of engineers focused on the Azure AI ecosystem.

The repository collects resources on Azure OpenAI services, LLM fundamentals (scaling laws, fine-tuning, alignment), retrieval-augmented generation (RAG), agent frameworks, orchestration, MCP, A2A protocols, evaluation, and LLMOps. Each entry gets a brief description and a date corresponding to its first commit, publication date, or paper release.

The README describes the collection as comprehensive and curated, with monthly updates and pre-update candidate tracking in the issue tracker. It is aimed at practitioners who want a single browsable reference rather than a textbook or a course.

The Three-Era Navigation Structure

The README organizes the collection through a layer-and-era model it calls Propedia-style. Three layers reflect how the field has developed over time.

The Weights layer covers the 2022 to 2023 period and addresses parametric knowledge: pretraining, scaling laws, fine-tuning, RLHF, alignment, and instruction following. The jump-to links from this layer point to sections on the large language model landscape, foundation model providers, and model training and optimization.

The Context layer covers 2023 to 2024 and focuses on what the model sees at inference time: prompting, chain-of-thought reasoning, RAG, memory, long context, and knowledge injection. Its sections include prompt engineering, Azure AI Search, RAG best practices, and context-window management.

The Agentic Engineering layer covers 2025 to 2026 and addresses how agents act, self-correct, and coordinate in the real world. Topics include harness engineering, loop engineering, graph engineering, function calling, tool ecosystems, MCP, skills, multi-agent systems, A2A protocols, and agent infrastructure.

Each layer in the README table links directly to the relevant sections in the section/ directory, making it a navigation index rather than a standalone document.

Cloning and Navigating the Collection

The repository is a set of Markdown files. Clone it to browse locally or read it directly on GitHub:

git clone https://github.com/kimtth/azure-openai-llm-notes

The top-level README is the primary navigation surface. It contains the three-era table with jump links that go to specific anchor points in the section/ directory files. The section/ directory holds the detailed content: separate Markdown files for applications (applications.md), Azure services (azure.md), models and research (models_research.md), best practices (best_practices.md), tools (tools_extra.md), and high-star LLM applications (x_llm_apps.md).

The code/ directory contains Python files. The files/ directory holds other assets. The .agent/ and .github/ directories are infrastructure for the repository itself.

For a reader who wants to track updates, the README states that candidate entries are tracked in the issue tracker before the monthly update cycle runs, which provides a preview of what is coming.

How Entries Are Organized Within Each Section

Each resource entry in the section files follows a consistent pattern: a link to the resource, a brief description of what it covers, and a chronological date. The README describes the date as the first commit date, the publication date, or the paper release date, depending on the resource type.

This chronological ordering means a reader can scan a section to see how the field has developed over time. An entry dated 2023 and an entry dated 2026 in the same RAG section will show how the tooling and approach evolved, even without a curated narrative.

The README also notes a separate section for popular LLM applications with GitHub stars at or above 1000, kept in x_llm_apps.md. This is the one place where stars are used as an organizational criterion (to filter into a separate list) rather than as a quality judgement within the main sections.

Monthly updates mean the collection ages more gracefully than a one-time snapshot, but readers should verify that linked resources are still available and current, since external links can become stale even when the collection itself is maintained.

What the Collection Leaves Out

azure-openai-llm-notes is a link-and-summary collection, not an explanatory resource. It does not walk the reader through how to configure Azure OpenAI, how to set up RAG with Azure AI Search, or how to build an agent from scratch. Readers who need step-by-step guidance need to follow the linked resources into their own documentation.

The collection is Azure-focused. The README has a separate companion repository (awesome-azure-openai-copilot) that focuses specifically on Azure and Microsoft products. Resources that cover only non-Azure providers may be underrepresented.

The repository has no licence declaration. The README does not state terms for using, redistributing, or building on the curation itself. Teams that want to reuse the structure or content of the collection should contact the author first.

The last push to the repository was on 2026-08-31. The project is not archived. Monthly updates mean the collection tracks the pace of Azure OpenAI platform changes, but any specific entry may describe a feature, service, or API that has since changed.

Compared with Microsoft's Official Documentation

Microsoft publishes its own Azure OpenAI documentation, API references, and code samples through Microsoft Learn and Azure docs. Those resources are authoritative, version-controlled by Microsoft, and structured for step-by-step onboarding.

The practical difference is scope and discovery. Microsoft's official documentation covers Azure OpenAI services in depth but does not curate community tools, research papers, open-source agent frameworks, or third-party RAG implementations. azure-openai-llm-notes covers that wider surface by collecting and briefly describing external resources alongside Microsoft's own materials.

A practitioner building an Azure-native AI application will likely use both. The official documentation handles implementation details; this collection handles discovery of the tooling and research landscape around the platform.

Maintenance and No-Licence Considerations

The README states that the collection receives monthly updates, with candidate entries tracked in the issue tracker before each cycle. The last push was on 2026-08-31, and the project is not archived, consistent with an active monthly update schedule.

The repository has no licence file and no licence identifier in the repository metadata. This is relevant for any team that wants to mirror, repackage, or build a product on top of the curation. The absence of a licence does not grant permission to use the work beyond reading it; default copyright applies in most jurisdictions.

The code/ directory contains Python files, but the README does not document their contents or purpose. Teams interested in the code should examine it directly after cloning.

Editorial conclusion

azure-openai-llm-notes is a practical starting point for engineers who need a maintained, browsable index of Azure OpenAI, LLM, RAG, and agentic engineering resources with brief descriptions. It is not a tutorial, a course, or a hands-on guide. Teams looking for executable code samples should check the linked Microsoft documentation and the code/ directory in the repository. No licence is stated, so redistribution or reuse of the curation itself should be discussed with the author before proceeding.

Frequently asked questions

Official sources

1. Issues
2. kimtth/azure-openai-llm-notes on GitHub
3. Project website
4. README