Part 6: Building AI applications
Build real products on top of models: calling LLMs, prompting, embeddings, RAG, agents and tools, multimodal apps, evaluation, security and responsible AI.
Key terms in this part (57)
- Generative AI
- Agentic AI
- Agent
- Agent loop
- Agent memory
- Bias (in data)
- Chain of thought
- Chunking
- Citation
- Computer use
- Context packing
- Cosine similarity
- Durable workflow
- Embedding
- Embedding model
- Evaluation set (eval dataset)
- Excessive agency
- Federated learning
- Few-shot prompting
- Function calling (tool calling)
- GraphRAG
- Guardrails
- Hallucination check
- Handoff
- Human in the loop
- Hybrid search
- LLM as judge
- MCP (Model Context Protocol)
- Metadata filter
- Model selection
- + 27 more in the glossary
- Call a modelAPI or local, structured output
- Add knowledgeEmbeddings and RAG
- Add actionsTools and agents with approval
- Make it trustworthyEvaluation, security, responsibility
Talking to models
- Calling an LLM from your application
Call a language model from Python and C#, request structured JSON, validate it, handle failures, stream answers and control cost.
- Prompt engineering that works in production
Write clear prompts with roles, rules, delimiters, examples and output formats, keep prompts versioned, and test them with a small evaluation set.
Knowledge: embeddings and RAG
- Embeddings and vector search
Explain embeddings, find similar text by cosine similarity, and store and search vectors in a vector database from Python and C#.
- RAG: answering from your own documents
Build a retrieval-augmented generation (RAG) system that checks permissions, retrieves evidence, answers with citations, says "I don't know" when evidence is missing, and is measured with an evaluation set.
- Advanced RAG: chunking, hybrid search, re-ranking and evaluation
Improve RAG quality with better chunking, hybrid keyword-plus-vector search, reciprocal rank fusion, re-ranking, query rewriting and metadata filters, and measure each change.
Actions: tools and agents
- Tools and agents with human approval
Let a model call tools safely, design an agent workflow with explicit state, permission checks, human approval, retries, idempotency and an audit trail, in Python and C#.
- AI frameworks and MCP: what to use in Python and .NET
Know what LangChain, LangGraph, Microsoft.Extensions.AI, Semantic Kernel, Microsoft Agent Framework and the Model Context Protocol do, and when plain code is better.
- Multimodal AI apps: images, documents and voice
Send images and documents to vision-capable models, build image-and-text search, and add speech input and output to an application.
Quality and trust
- Evaluating AI applications
Build an evaluation set, write automatic checks, use an LLM as a judge carefully, detect hallucinations, and compare versions with A/B tests.
- Security for AI applications
Recognise the main risks to LLM applications (prompt injection, data leakage, excessive agency and others), apply layered defences, and redact personal data before it reaches a model.
- Responsible AI: fairness, privacy, transparency and oversight
Measure fairness between groups, protect privacy, explain decisions, keep people in control, and document models with model cards and risk registers.
Advanced: the whole picture
- "Generative AI and Agentic AI: the whole picture"
Explain in simple words what generative AI and agentic AI are, how they differ and work together, and decide when a task needs one, the other, both, or neither.