← All paths
AI agents in production
Agents that run real tasks, not demo chatbots: tool calling, the Claude and OpenAI agent SDKs, MCP, context engineering, RAG over real documents, memory, multi-agent orchestration, evals, security and sandboxing, cost, and operating agents in production.
- Route
- 17 guides · 928 lessons
- Level
- Beginner to advanced
Agent Fundamentals and Tool CallingLearn from scratch how an AI model goes from answering to acting: build an agent with tools, a decision loop, and error handling in Python.Start →
Agents with Real SDKs: Claude Agent SDK and OpenAI Agents SDKBuild the same agent with Anthropic's SDK and OpenAI's, with human approval and cost control, and use real data to choose the one that fits you.Open →
MCP Deep DiveExpose your agent's tools over MCP by building a server and a client by hand, so you understand the protocol itself and not just an SDK.Open →
Context EngineeringDecide with discipline what goes into a model's context window: budget tokens, order and compress information, and measure the effect of every change.Open →- Embeddings Deep DiveTurn text into vectors your team can search: compare embedding models, pick the right distance metric, chunk with care and measure retrieval quality before it reaches production.Open →
- Vector Database FundamentalsUnderstand how vector databases work inside, why a RAG system needs one, and how to choose the right one for your team's project.Open →
Production RAG and Document IngestionGive your agent reliable access to real documents: clean them, chunk them with care, search them like any other tool, and cite the source in every answer.Open →- Advanced RAG TechniquesTake a RAG system past the basics: hybrid search, reranking, query rewriting and evaluation, so your team can see which change actually improves the answers.Open →
Agent Memory and StateGive your agent a memory that survives the end of a session, so it remembers each user, updates what changed, and forgets when asked.Open →
LangChain and LangGraph: From Chains to AgentsBuild production-ready AI agents and workflows with memory, human approval, and monitoring that go well beyond a chatbot.Open →
Multi-Agent OrchestrationDecide when it's worth splitting work across several agents and, when it is, coordinate them by hand in Python without relying on any framework.Open →
Long-Running Agents: The HarnessBuild the system that lets an agent work for hours or days across many sessions without losing track, repeating actions, or calling untested work done.Open →
Evaluation FrameworksFind out with evidence whether your LLM assistant works: read its conversations, find where it fails, measure each failure with trustworthy evaluators, and decide when it's ready to launch.Open →
Agent Security and SandboxingMake sure an agent with real tools can't be tricked into deleting, spending, or leaking data, even when the model or the person typing to it is hostile.Open →
Cost Optimization and CachingCut your LLM system's bill with savings you can measure and defend: shorter prompts, provider caching, your own cache, the right model, and batch processing.Open →
Computer Use and Browser AgentsBuild an agent that works in systems with no API the way a person would: it reads the screen, clicks, types, checks the result, and asks before anything irreversible.Open →
Agents in ProductionTake an AI agent to production and run it with a trace of every step, measured cost and latency, and a regression check before every change.Open →
Your team designs, evaluates and operates agents that do real work, with cost and security under control.