Vlad Annenkov
Hi, I’m Vlad Annenkov, a software engineer at JetBrains, based in Madrid. I work on Rider, the .NET IDE, where I build the tools AI agents use to work with real code.
What I’m working on
I designed and own Rider’s shared infrastructure for agent skills, hooks and MCP integrations, the layer other Rider teams build their AI features on. I’m also Rider’s technical point of contact with JetBrains’ central AI Assistant team.
Most of this work comes from one idea: an IDE already understands code better than an agent that reads and edits it as text, so the agent should use what the IDE knows instead of guessing. Two things I’ve built along these lines:
- A refactoring skill that lets agents call Rider’s refactoring engine instead of rewriting code by hand. On 15 C# refactoring tasks, median task time dropped by 83% and cost per solved task by 64%.
- Quality-check hooks that check every change an agent makes with Rider’s inspections. Errors stop the agent from marking the task as done; warnings go back to it as feedback.
Before that
Before agents, I spent most of my career teaching editors to understand code. On Rider Core (2023–2026), I built the code completion infrastructure for Rider’s new editor engine from scratch and worked on how AI and standard completion fit together. I also shipped the Type Dependency Diagram, a top-voted feature request that I first built as an MVP during my JetBrains internship.
Earlier, I worked on a cloud IDE at 1C, building support for mutually embedded languages, and briefly as a backend developer at Tinkoff. I studied Computer Science at HSE University.
This blog
I write about AI agents in the IDE, developer tools, and what I learn from evaluating agents on real engineering tasks.