Overview
AI Dev Conf 2026 Autumn is a conference on agentic development and the practical use of AI in software engineering. Talks take place on 26 November 2026 and hands-on workshops on 28 November 2026. The event is supported by JUG Ru Group and features two tracks.
The program focuses on how teams can adopt AI, keep company data safe, and measure real results instead of hype. It covers agent-based systems, engineering context such as project memory, skills and MCP, product metrics, evaluation of agent products, and observability of AI agents.
Who should attend
- Managers, engineering leaders and infrastructure teams who want to introduce AI with measurable outcomes and control data security.
- Team leads, product managers and tech leads who need to manage development when part of the code is written by agents.
- Developers, DevOps engineers and testers who want to build agent systems and apply practical tools and practices.
Agenda
- 11:00 - 12:45: Observability of AI agents with Langfuse. A workshop on tracing LLM calls, collecting metrics, evaluating answer quality and diagnosing agents.
- 13:00 - 14:30: Homo agenticus v2. Transferring agent context from product owners to development teams.
- 14:45 - 16:15: How to build your own QAI agent for task testing and root-cause analysis. Speaker: Alexander Volkov.
- 15:00 - 15:30: Margin outside the IDE: what breaks when AI speeds up everything except promises. Speaker: Grigory Dobryakov.
- 15:45 - 16:30: Shared AI agents: from personal chats to collaborative development. Speaker: Denis Slashchev.
- 16:45 - 17:30: Hypervelocity Engineering: building production-ready solutions. Speaker: Pavel Golubev.
- 16:45 - 17:30: How to cover the entire SDLC with an agent. Speaker: Pavel Fedotovsky.
- 17:45 - 18:30: Evals for agent products: what to cover and when to stop. Speaker: Evgeny Sergeev.
- 18:45 - 19:30: Drunk on AI Speed: what happens when management decides to point agents at production code. Speaker: Alexander Volochnev.
- 19:40 - 20:30: Round table "Lost in Semantics" on semantic layers, ownership and trusting agents with data.
- 19:35 - 19:40: Closing and summary of results.
Other talks include a review of AI agent pilots in Russian fintech, telecom, industry and retail companies, an account of scaling AI use from 10 to about a hundred engineers, and a discussion of the state of AI in SDLC.
Format
The conference has two tracks and combines talks with interactive workshops. Registration is available on the official website, and an online option is mentioned on the page. The exact venue is not stated in the visible content.
Speakers
The lineup includes engineers and leaders from Yandex, Microsoft, CodeRabbit, Raft, MWS, Kaspersky, RuStore, T-Bank, Sber and VK, along with independent experts in AI engineering and quality assurance.




