Agents Week

powered by the Yandex School of Data Analysis

This intensive is designed for anyone planning to implement agent systems in real products or infrastructure, or build personal assistants:

• ML engineers, backend developers, and researchers who want to move from isolated LLM prompts and prototypes to full agent systems.
• Technical university students with strong coding skills eager to understand AI agents.

Practical assignments require familiarity with Python, APIs, and core LLM concepts (tokens, inference, context windows). If you are new to LLMs, we recommend reviewing the basics beforehand.

• How modern AI agents work: from LLMs and prompts to the thought—action—observation loop, tools, memory, and agent architectures.

 

• Designing and configuring agent behavior: reasoning, orchestration of multi-agent systems, and guardrails, with real engineering constraints.

 

• Evaluating AI agents: сore approaches to assessing solutions and checking the reliability and stability of agent behavior.

 

• Production engineering principles for AI agents: setting up monitoring, scaling, and maintaining systems under real workloads.

• the intensive is available in English and Russian. Presentations and tasks are in English, while speakers present in Russian. Subtitles or Gemini AI dubbing can be enabled.

 

• Video sessions will be streamed in the evening as pre-recorded broadcasts on YouTube. After the intensive, we will host a Q&A session where participants can ask questions.

 

• While video streams are open to everyone, earning a certificate requires completing the selection tasks and the final assignment.

 

• If you pass the selection stage, access to the final assignment will appear in your Personal Account.