openai demo

pdebruin · Sep 10, 2024 · 98db0dc · 98db0dc
1 parent d9bccde
commit 98db0dc
Show file tree

Hide file tree

Showing 7 changed files with 24 additions and 8 deletions.
diff --git a/_posts/2023-12-01-csharp-for-beginners.md b/_posts/2023-12-01-csharp-for-beginners.md
@@ -7,11 +7,9 @@ tags:
   - Devops
 ---
 
-todo
+Interested in learning c# (or how to create dev videos)? Watch the [c# for beginners series](https://learn.microsoft.com/shows/csharp-for-beginners/). 
 
 ![img](../assets/images/2023-12-01-csharp-for-beginners.png)
 
-https://learn.microsoft.com/en-us/shows/csharp-for-beginners/
-
 Thanks for reading! :-)
 
diff --git a/_posts/2023-12-22-mslearn.link.md b/_posts/2023-12-22-mslearn.link.md
@@ -7,11 +7,15 @@ tags:
   - Devops
 ---
 
-todo
+In January 2022 I moved from Microsoft Netherlands to Microsoft Corp to work on Microsoft Learn. At the time our team started to run various programs to grow contributors and contributions to Learn content, initially mostly documentation. After various reorganizations, I work on slightly different things, but still for Microsoft Learn. I like the domain, since it allows me to work with technology and experiment with free knowledge sharing and thought leadership. 
 
-![img](../assets/images/2023-12-22-mslearn.link.png)
+Sometimes I also see things that can be improved :-) which I then mostly discuss internally. In this case I was unhappy with the experience for Microsoft Learn users, who can land on a number of different pages that eventually all point to the same documentation, trainings, etc. As a result, I wanted to experiment with providing a flat list of unique entry points to Learn content that is searchable and that takes away all marketing and landing pages. 
+
+The result is [mslearn.link](https://mslearn.link).
 
-https://mslearn.link
+Let me know what you think. Good? Bad? Don't care? Other? 
+
+![img](../assets/images/2023-12-22-mslearn.link.png)
 
 Thanks for reading! :-)
 
diff --git a/_posts/2024-09-13-ai-samples.md b/_posts/2024-09-13-ai-samples.md
@@ -6,8 +6,22 @@ tags:
   - Devops
 ---
 
-https://github.com/Azure-Samples/azure-search-openai-demo
+Starting from scratch can be daunting. I prefer to reuse a sample of some sort and then customize to my needs. This means you may have a bit of a learning curve to understand both the sample as well as the technology it uses. This post should help.
 
-![img](../assets/images/2024-09-06-rag-hack.png)
+The sample: [Azure Search OpenAI demo](https://github.com/Azure-Samples/azure-search-openai-demo)
+
+I wanted to write about this sample for some time, because it includes a number of things that I find fascinating in the new reality with artificial intelligence everywhere: It uses both [Azure OpenAI service](https://learn.microsoft.com/azure/ai-services/openai/overview), Azure's flavor of OpenAI's language models, as well as [Azure AI Search](https://learn.microsoft.com/azure/search/search-what-is-azure-search), which has been around longer. It also provides scenario data from Contoso Electronics, which demonstrates how you can make the sample work with your data. Lastly, the sample is very thorough in the sense that it contains bicep scripts to provision Azure services through [Azure Developer cli AZD](../azd), source code for the apps, documentation for various scenarios, and tests. 
+
+Since the sample is rather thorough, it has quite a number of dependencies. To get started without worrying about those, you should use GitHub Codespaces if you can. Deploying the app is done through ``` azd up ```.
+
+This will additionally result in an [Azure app service](https://learn.microsoft.com/azure/app-service/overview) that hosts the app, an Azure storage account](https://learn.microsoft.com/azure/storage/blobs/storage-blobs-overview) to store the scenario pdf files, and [Azure AI Document Intelligence service](https://learn.microsoft.com/azure/ai-services/document-intelligence/overview) to analyze them. 
+
+![img](../assets/images/2024-09-13-ai-samples1.png)
+
+Once the deployment is finished, you can access the app through the url of the app-backend. You can then interact with the solution, for instance by asking it about health plans. Finally the chat response contains citations, which you can click on to see the source pdf file. 
+
+![img](../assets/images/2024-09-13-ai-samples2.png)
+
+In following posts, let's look at those Azure services to understand what they do and whether you need them in your app. 
 
 Thanks for reading! :-)
diff --git a/assets/images/2023-12-01-csharp-for-beginners.png b/assets/images/2023-12-01-csharp-for-beginners.png
diff --git a/assets/images/2023-12-22-mslearn.link.png b/assets/images/2023-12-22-mslearn.link.png
diff --git a/assets/images/2024-09-13-ai-samples1.png b/assets/images/2024-09-13-ai-samples1.png
diff --git a/assets/images/2024-09-13-ai-samples2.png b/assets/images/2024-09-13-ai-samples2.png