Skip to content
View kurochenko's full-sized avatar
πŸ›³οΈ
πŸ›³οΈ

Block or report kurochenko

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
kurochenko/README.md

Andrej Kurochenko

Senior developer & wannabe builder.

Projects

Mend β€” Autonomous code review for GitLab and GitHub. Runs on your own machine with your coding agent of choice. GitLab is supported today, with GitHub support coming soon.

cookt.io β€” Recipe organizer with AI. Paste any recipe format, get structured ingredients, steps, and chat with your recipes.

skillbook.dev β€” CLI tool for managing AI skills across projects and teams. Centralized library that works with any AI harness (OpenCode, Cursor, Claude Code).

pi.nvim β€” Neovim plugin that integrates the Pi coding agent into a side panel. Select code, send it to Pi with a keypress, and get responses back β€” with context placeholders, an action picker, and built-in prompts for explain, review, fix, test, and more.

living-spec β€” Stop your AI from making things up. Framework that captures hidden project knowledge β€” invariants, decisions, rules β€” in a format your LLM reads before it writes code. Catch contradictions early, auto-generate tests from specs, and make tribal knowledge explicit.

What I Do

Help teams integrate AI into their development workflow:

  • Automating bug fixing and code review pipelines
  • Streamlining development processes
  • Building tools that let developers focus on what matters

Case Studies

Production Incident Response: From Hours to Minutes

Problem: Production incidents required manual investigation across multiple systems β€” checking logs in CloudWatch, querying ALB logs in Athena, reviewing exceptions in Sentry, finding relevant code, and figuring out fixes. This took hours of senior engineer time.

Solution: Built AI agents that connect monitoring tools (CloudWatch, Sentry) to code repositories (GitLab). When an alert fires, the AI automatically pulls relevant logs, identifies the root cause, and creates a merge request with a proposed fix.

Result: What used to take hours of manual investigation now takes 5 minutes. Engineers review the AI-generated MR instead of hunting for the problem.

Read article β†’

Code Review Automation

AI pipeline that analyzes PRs, suggests fixes, and auto-generates merge requests for common issues.

Background

Started with backend (JVM, Node), moved through frontend web (React) and mobile (React Native), then cloud/serverless, Linux, and DevOps.

Lately: TypeScript for everything β€” backend, frontend, terminal UIs.

Pinned Loading

  1. tanstack-start-breadcrumbs-example tanstack-start-breadcrumbs-example Public

    Example project showcasing breadcrumbs in Tanstack Start with SSR and dynamic update of breadcrumbs on data change.

    TypeScript 24 2

  2. skillbook skillbook Public

    TypeScript 38 5