AI Engineering

34 articles on ai engineering.

• Request a response with structured output from an LLM using Microsoft.Extensions.AI in .NET — I used to hand-write the prompt, get back a paragraph, then trim and parse it into an object myself. Microsoft.Extensions.AI let me skip that entire middle step. Here's the difference, with a real example.
• Giving your AI agent a wallet: my Alfred moment — I gave my AI agent a payment credential and let it buy my Spider-Man tickets. Here is how Stripe's link-cli works, what Mastercard and Visa are shipping, and why your agent might soon have a better wallet than you.
• I Passed the CCA-F Exam on My First Try. Here Are the 5 Ideas That Did Most of the Work — I recently cleared the CCA-F certification. Here is what the exam actually rewards, the five ideas that carried me through every domain, and the recurring question traps you should be able to spot on sight.
• Microsoft Agent Framework: The Successor to Semantic Kernel and AutoGen — Microsoft folded Semantic Kernel and AutoGen into one SDK called Microsoft Agent Framework. Here's what it actually looks like in C# and Python, the agent and workflow primitives, and the orchestration patterns it ships with.
• The Agent2Agent (A2A) Protocol: Learning v1.0 by Building a Small Agent — A2A v1.0 shipped in March. I built a small reference agent on top of the official Python SDK and wrote down everything that clicked, plus what the SDK hasn't caught up to yet.
• Google Web MCP: making your website agent-ready — Google just dropped WebMCP, a browser API that lets websites tell AI agents what they can do. No more DOM scraping. Here is what it is, how to use it, and why it matters.
• Agents vs sub-agents: what they are and how they work together — Agents and sub-agents are structurally the same thing. The difference is who calls whom. Here is what that means in practice.
• How To Build a Web Scraping MCP Server with Python — Building a basic MCP server in Python that scrapes web pages and feeds clean content to LLMs.
• Understanding ADLC: The Agentic Development Life Cycle — The Agentic Development Life Cycle (ADLC) is a framework for how you build, evaluate, deploy, and operate AI agent systems. This post covers the phases with diagrams, code, and lessons from real projects.
• The State of AI Agents in 2026: Frameworks Tools and Best Practices — Where AI agent frameworks stand in early 2026 across Python and .NET, with notes on what actually works in production.
• Multi-Agent Systems with .NET 10 and Semantic Kernel Process Framework — Using Semantic Kernel's Process Framework in .NET 10 to wire up multi-agent workflows with steps, events, and process orchestration.
• Building a Zoho Mail RAG System with Python, FastAPI, and ChromaDB — A full-stack RAG system that searches your Zoho Mail inbox with natural language, built on FastAPI, ChromaDB, and GPT-4o-mini with conversation memory.
• Threat Intelligence Graph RAG with Neo4j and Ollama — A local, containerized Graph RAG system that queries MITRE ATT&CK data through natural language using Neo4j and Ollama.
• AI Agent Development with Anthropic Claude Agent SDK — Building AI agents with Anthropic's Claude Agent SDK in Python, covering tool use, agent loops, structured outputs, and multi-turn conversations.
• Advanced Agent Orchestration Patterns in Python — Supervisor hierarchies, tool-use strategies, error recovery, and human-in-the-loop checkpoints for Python agent systems that need to work reliably.
• Building AI Agents with .NET 10 and Microsoft.Extensions.AI — .NET 10's Microsoft.Extensions.AI library gives you IChatClient, middleware pipelines, tool calling, and DI patterns for building agents in C#.
• Multi-Agent Orchestration with AutoGen Studio — Using AutoGen Studio to build multi-agent systems visually, with team configuration, workflow design, and custom skills.
• Building AI Agents with OpenAI Agents SDK in Python — Building agents with OpenAI's Agents SDK in Python, covering Runner execution, handoffs between agents, guardrails, and the built-in tracing system.
• Python Agent Framework Comparison: LangGraph vs CrewAI vs AutoGen — Comparing LangGraph, CrewAI, and AutoGen side by side with architecture breakdowns, code, and guidance on when to pick which.
• Building AI Agents with .NET 9 and Semantic Kernel Agents — How to build AI agents in .NET 9 using Semantic Kernel's Agents framework, with ChatCompletionAgent, OpenAIAssistantAgent, and multi-agent collaboration.
• A .NET-Powered RAG Console Application with Ollama — How to wire up a RAG console app in .NET with Ollama running locally, and why .NET is a better fit for this than most people think.
• Multi-Agent Workflows with CrewAI and LangChain Integration — Combining CrewAI's crew-based orchestration with LangChain's tool ecosystem to build multi-agent workflows for automation tasks.
• Agent Orchestration with Microsoft AutoGen 0.4 — AutoGen 0.4 was rebuilt from scratch with an event-driven architecture and a new AgentChat API. This post walks through the team-based orchestration model and what changed from 0.2.
• Advanced RAG Patterns with Python and ChromaDB — Taking basic RAG further with smarter chunking, hybrid search, reranking, and metadata filtering in Python and ChromaDB.
• AI-Powered Applications with .NET Aspire and Semantic Kernel — Combining .NET Aspire and Semantic Kernel to build distributed AI apps with proper orchestration and observability built in.
• Multi-Agent Debate and Collaboration Systems with AutoGen — Building debate and collaboration systems in AutoGen with group chat orchestration, speaker selection, and code execution.
• Agent Orchestration with LangGraph in Python — How to build stateful AI agents with LangGraph, including graph construction, conditional routing, and state management across conversation turns.
• Building AI Agents with Semantic Kernel and .NET 8 — Building AI agents with Semantic Kernel in .NET 8, covering function calling, plugins, auto-invocation, and planning.
• Fine-Tuning Large Language Models with Python and Hugging Face — Fine-tuning LLMs with Hugging Face Transformers and LoRA (PEFT), from dataset prep through training and evaluation.
• Building Multi-Agent Systems with CrewAI and Python — Using CrewAI in Python to build multi-agent systems where agents have defined roles, goals, and coordinated workflows.
• Retrieval-Augmented Generation with LlamaIndex and Python — Using LlamaIndex and Python to build a RAG system that grounds LLM answers in your own documents with vector stores and query engines.
• Building Multi-Agent Conversations with Microsoft AutoGen — Setting up multi-agent conversations in AutoGen 0.2 where agents collaborate and debate to solve problems together.
• Getting Started with Semantic Kernel for .NET Developers — Getting started with Semantic Kernel 1.0 in .NET 8: kernel setup, plugins, planners, and basic AI app patterns.
• Building Your First AI Chatbot with Python and LangChain — Building a chatbot with Python and LangChain v0.1 that maintains conversation memory using the OpenAI API.