Alexey Seleznev released a free video course on R for developing AI tools

Alexey Seleznev has published a free video course on R focused on creating AI tools, chat interfaces, and multi-agent systems. The course is designed as a practical roadmap for those who already write in R and want to integrate LLMs into their workflows, rather than just using web chats.

What's inside the course

The main idea of the course is to demonstrate that the R ecosystem is suitable not only for classical analytics and visualization, but also for full-fledged development of AI applications. The author focuses not on a collection of disparate tricks, but on a cohesive stack of tools: `ellmer` for working with models, `mcptools` for MCP, `ragnar` for RAG, `shinychat` and `querychat` for interfaces, and `mini007` for agent scenarios. As a result, the course looks not like a review of novelties, but as an attempt to assemble a ready-made trajectory for R developers entering applied AI.

"This course is not about abstract examples and toy demos".

It is also important that the materials are available for free in the format of an online book with video lectures, notes, and code examples. This lowers the barrier to entry for those who have long worked in R but still viewed modern AI frameworks as Python territory. In this form, the course solves two tasks at once: it provides a learning structure and shows which packages to use right now, without independently assembling a stack from dozens of repositories and articles.

Seven practical modules The program consists of seven consecutive lectures.

The author first covers basic work with LLMs through the `ellmer` package: configuring API keys, creating chats, extracting structured data from text, and building a simple interface. Then the course moves to more applied matters — deploying a model to a Telegram bot, managing user context, and preserving chat history between sessions.

• Connecting different LLM providers directly from R Creating AI chats and web interfaces based on Shiny Running an MCP server and MCP client for working with external tools Building a RAG system with embeddings, DuckDB, and hybrid search Developing multi-agent scenarios with R code generation and execution In the second half of the course, the focus shifts to more complex architectures. One module is devoted to MCP as a way to connect models with data and functions, including integration with tools like Claude Desktop. Another addresses the RAG approach: embeddings, vector storage on DuckDB, and answers based on your own documentation. Separate lessons cover configuring the `shinychat` interface, working with `querychat`, which translates natural language queries into SQL, and multi-agent systems where AI not only answers but also coordinates actions, evaluates result quality, and executes code.

Who the course is for

According to the author's description, the course is primarily aimed at data analysts, R developers, and those who already write R code confidently. This is not introductory material for absolute beginners: without basic language syntax and an understanding of how to work with packages, the pace will likely be too fast. However, for practitioners who want to quickly move from "tried a chatbot" to a working prototype, the program looks quite practical.

The strength of the course is that it covers several working scenarios that are usually studied in parts. Here in one place you will find LLM APIs, bots, web interfaces, data access through MCP, search across your own knowledge base, and agent patterns. For corporate teams this is especially useful: an analyst or BI developer can avoid changing the main language and stack, and add AI features directly to existing R projects, dashboards, and internal services.

The free format also matters: such a course can be used as an internal entry point for a team without a separate training budget.

What this means R remains a niche language compared to

Python in AI development, but courses like this show that the gap is rapidly narrowing at the level of applied tools. For the Russian-speaking community, this is a ready-made, structured way to enter LLMs, MCP, RAG, and agent systems without leaving the familiar ecosystem.