AutoGen Conversational Agents
Microsoft AutoGen v0.4 - event-driven multi-agent runtime, AgentChat teams, code execution, and production patterns for conversational AI systems.
Microsoft AutoGen v0.4 - event-driven multi-agent runtime, AgentChat teams, code execution, and production patterns for conversational AI systems.
CrewAI in production - agents, tasks, crews, memory systems, Flows, and deep-dive patterns for role-based multi-agent pipelines.
Comprehensive comparison of LangGraph, CrewAI, AutoGen, LlamaIndex, and raw API across 12 production dimensions - with decision flowchart and real case studies.
replaced
Graph-based stateful agent orchestration with LangGraph - StateGraph, typed state, nodes, conditional edges, checkpointing, and human-in-the-loop.
LlamaIndex's document-centric agent framework - VectorStoreIndex, QueryEngine, FunctionCallingAgent, and the Workflow event-driven orchestration model.
LangChain, LangGraph, LlamaIndex, CrewAI, AutoGen, raw API - an honest comparison with production lessons.
12 hard-won lessons from deploying agentic systems at scale - each with a war story, a principle, and a code pattern you can use today.
Building production agents with just the Anthropic SDK - the agentic loop, tool handling, context management, cost tracking, and a complete 200-line implementation.
The framework vs raw API decision for agents - what abstractions cost, what they provide, and a decision tree based on your actual requirements.