AI Development in .NET Microsoft.Extensions.AI

AgentCon Istanbul - 07.02.2026
AI Development in .NET Microsoft.Extensions.AI

Transcript

1. AI Development in .NET Mert Metin #AgentConIstanbul Microsoft.Extensions.AI
2. Agenda Basic Concepts of GenAI Exploring the .NET Package - Microsoft.Extensions.AI
3. Brief Recap of the GenAI Concepts Generative AI, LLM, Prompt, Roles, RAG ,Agent, Token, Embedding, Tools
4. Generative AI Subset of AI Creates new content such as text, code, audio, image based on user prompts and learned data.
5. Large Language Model LLMs are optimized for complex tasks, trained on huge datasets, and capable of understanding and interpreting human language.
6. Prompt An input that is used to communicate with the model to generate desired output. Prompt Engineering is the practice of designing effective prompts. Prompts must be; clear, specific, goal oriented, have contextual information
7. Agents They are capable of reasoning, planning, and interacting with their environment, using generative AI to perform tasks autonomously.
8. System Specify “How will the model behave?” Roles User Prompt and query by user. Asistant Model responds using user’s prompt. It is used to ensure focus on a specific task or behavior pattern.
9. Tool Functions provided to LLMs must have clear objectives. Tools enable agents to perform actions such as web search, retrieving information from external sources and API calls.
10. Token Meaningful unit processed by LLM. Token is not only a word. tokenized by gpt-4o
11. Embedding - Vector Database Embedding is an operation that converts text, images into numerical data (vectors) that machines can understand. A vector database stores numerical data as vectors, enabling efficient similarity search
12. Retrieval Augmented Generation (RAG) Enhances LLMs with external data sources (databases, documents, APIs). Improves response relevance and accuracy.
13. Exploring the .NET Package Microsoft.Extensions.AI Provides utilities for working with generative AI components.
14. Why Support for switching between different LLM providers Manageable, Simple, and Rapid Development Provide implementations for dependency injection, caching, telemetry and other cross-cutting concerns. Implement chat and embedding features using any LLM provider.
15. Use Cases Product description for e-commerce domain Content summarization from long articles Easy to categorize for news feeds Customer review analysis Call center assistant Internal frequently asked question assistant
16. Chat Feel like "using chat gpt, gemini" with abstraction
17. IChatClient Connecting to LLM providers such as Open AI, Ollama, Azure Open AI
18. Microsoft.Extensions.AI.OpenAI LLM provider for OpenAI Working With Model Providers
19. The eaisest way to use the Ollama in .NET. Useful for working on local machine. Working With Model Providers Note: Microsoft.Extensions.AI.Ollama package is deprecated. OllamaSharp
20. GetResponseAsync Sends a user chat text message and returns the response messages. Return type is ChatResponse which represents the response to a chat request.
21. GetStreamingResponseAsync Sends chat messages and streams the response. Return type is IAsyncEnumerable<ChatResponseUpdate> which provides a stream of updated chat response.
22. Case - Sentiment Analysis Customer’s order reviews of the restaurant
23. Output - GetResponseAsync
24. Output - GetStreamingResponseAsync
25. Conversation Application needs to keep conversation context. Case - git commit generator
26. Conversation Case - git commit generator ChatRole.System specifies “How will the model behave?” List<ChatMessage> Initialize the with a prompt of this role.
27. Conversation Each user’s prompt must set to Case - git commit generator ChatRole.User. ChatRole.Assistant AddMessages role is set by
29. History and Roles often represents the history of whole chat messages that are part of the conversation. Chat result List<ChatMessage>
30. Tool Calling - set method name as delegate type UseFunctionInvocation
31. Tool - Get DollarBanknoteBuying()
32. Output Usage Informations Response
33. Embeddings GenerateEmbeddingAsync has more detailed information about embeddings and Vector collection is also included. GenerateEmbeddingVectorAsync returns the vector collection.
34. Case - RelatedNews
35. Output
36. Evaluation Microsoft.Extensions.AI.Evaluation provides core abstractions to evaluate the quality of AI applications.
37. Quality Microsoft.Extensions.AI.Evaluation.Quality Provides evaluators for quality-based metrics Relevance Truth Completeness Fluency Coherence Retrieval Equivalence Groundedness.
38. Single vs Multiple Evaluator
39. Case - Testing Capital City
40. Provide context to evaluators that need it. Multiple evaluators using CompositeEvaluator
41. Evaluate the response Result may affect by different LLMs. Chat messages and model’s responses are essential part of the EvaluateAysnc
42. NumericMetric’s value between 0(unknown) and 6 (exceptional) Interpretations of evaluations
43. Reporting Microsoft.Extensions.AI.Evaluation.Reporting provides the support to cache LLM responses, store evaluation results and generate reports from that data.
44. Microsoft.Extensions.AI.Evaluation.Reporting
47. Dependency Injection Inject as ChatClient, EmbeddingClient OpenAI Ollama
48. Dependency Injection Call from API Controller
49. Dependency Injection Using Keyed Services
50. Caching Caching Alternatives: Memory or Redis
51. Caching UseDistributedCache pipeline for caching in IChatClient
52. Redis Cache
53. Open Telemetry
54. Jaeger
56. thanks for listening