LLMs & Autonomous Agents Tutorials

Learn LLMs & Agents: where to start

Large language models (LLMs) are AI systems trained on huge amounts of text that can understand and generate language — answering questions, summarizing documents, writing and explaining code, and following instructions given in plain words. An AI agent goes one step further: it uses an LLM as its reasoning core, plans a sequence of steps, calls tools such as search engines, calculators, databases, or APIs, and works through multi-step tasks with limited human input. The tutorials in this category cover both halves of that picture: how to prompt models and call them through their APIs, how to ground them in your own data with retrieval (RAG), and how to build agents with popular frameworks. The material is aimed at developers who already write some code — usually Python — and want to build applications on top of models instead of training models themselves.

Start by using a model directly: experiment in a playground or through an API, and learn how prompts, context windows, and settings such as temperature change the output. Then build a small chatbot that holds a conversation with short-term memory, and learn how applications count tokens and keep costs under control. Next comes retrieval-augmented generation — embedding your documents, storing them in a vector database, and letting the model answer from your own sources. After that, add tool calling, where the model requests functions that your code executes; that is the bridge from chatbot to agent. The advanced tutorials then cover agent frameworks, multi-step planning, and the unglamorous essentials: evaluating answers, logging behavior, and adding guardrails so your system fails safely instead of failing confidently.

What you can build with LLMs & Agents

• Support and FAQ chatbots that answer from your own documentation using retrieval, with sources you can check
• Document tools that summarize long reports, extract names, dates, and figures, and answer questions over PDFs
• Workflow agents that triage incoming messages, draft replies, and update a spreadsheet or database through tools
• Research and coding assistants that break a goal into steps, search or call APIs, and assemble a result
• Content pipelines that classify, tag, translate, and rewrite large volumes of text consistently

Frequently asked questions

Do I need machine-learning experience to build with LLMs?

No. Most application development happens through APIs: you send text in and receive text out, using ordinary programming skills. Python basics, comfort with JSON and HTTP requests, and careful prompt design will carry you through chatbots, RAG, and tool calling. Deeper machine-learning knowledge only becomes necessary if you want to fine-tune models or run and optimize them on your own hardware.

What is the difference between a chatbot and an agent?

A chatbot mainly converses: it takes your message and replies, sometimes with memory of the conversation. An agent is built to pursue a goal across several steps — it can decide which tool to use, act on the result, check its own progress, and continue until the task is done or it needs your input. In practice the two blend together, and many agents are chatbots with tools and a planning loop added.

How much does it cost to experiment with LLMs?

Less than most people expect at learning scale. Several providers include a free tier or starter credit, and API pricing is charged per token, so short experiments with smaller models cost very little; costs grow with longer inputs, larger models, and production traffic. You can also run open-weight models locally at no per-call cost if your computer has enough memory, which is a good way to practice without a bill.