Backend Engineer (Go/Golang) building high-load distributed systems at one of the world's largest classified platforms. AI Enthusiast.

Joined March 2020
I landed at #6 on @hackernoon’s Software Engineering writers list 🎉 My latest article is about guardrails for AI coding agents in Go: turning recurring review feedback into checks, then proving those checks catch real failures. What’s one guardrail you’d add?
1
2
25
I built Pumlpad: a free, open-source PlantUML editor for macOS. Native UI, live preview, fully offline. Java and PlantUML are bundled, so you can just open a .puml file and start editing github.com/krus210/pumlpad
1
21
Publish what your company learns! I made a new site --> businessblogging.tech
4
2
15
427
On September 10, I facilitated a system design kata: 15 participants in 5 teams designing a gamification engine for Avito. We assessed user scenarios, diagrams, reliability, data models, APIs, and load estimates. Congrats to the winning team!
2
7
Хочу поделиться своим Telegram-каналом AI Visual Digest. Собираю там статьи и видео про AI и разработку, которые мне интересны. Коротко описываю основные идеи и делаю визуальные конспекты со схемами и выводами. Ещё выпускаю дайджесты. t.me/ai_visual_digest
1
13
I helped organize Avito Go&Grill in St. Petersburg and facilitated a system design kata: build a festival meetup service for introverts. Teams tackled 50K users, EDGE, offline sync, matching and consent-based location sharing. Great fun. They nailed it! #Golang #SystemDesign
25
I benchmarked archspec, my open-source library, in 12 runs across 6 LLMs, with and without it, including GLM-5 and MiniMax-M2.7. Plan: 6.3 → 7.9 Checklist: 66% → 89% Benchmark: twilight-shadow-2321.staskor… Repo: github.com/krus210/archspec
1
23
I spoke at The Top Voices about taking AI from vibe coding to production. AI agents draft code fast. The hard part is context, tests, observability, and ownership. Codegen is no longer the bottleneck. Production engineering is. youtu.be/OlTTKZ24iek?t=1025
2
2
41
Release day: Clapet est en ligne! clapet.app/ J'ai construit le tool de System Design que j'aurais aimé avoir pour préparer mes entretiens. Clapet te guide pas à pas pour construire une archi au lieu de te laisser seul devant un canvas vide a la Excalidraw.
90
227
20
2,453
294,131
今週GitHubで急上昇したAIリポジトリ10選: 1. headroom LLMに渡す前にログ・ファイル・RAGチャンクを圧縮し、トークン消費を60〜95%削減するツール。ライブラリ・プロキシ・MCPサーバーの3つの形で組み込める。 github.com/chopratejas/headr… 2. Agent-Reach Twitter・Reddit・YouTube・GitHubなど主要プラットフォームをAIエージェントに横断検索させるツール。APIキー不要でCLI1本から動く。 github.com/Panniantong/Agent… 3. agent-skills(addyosmani) コーディングエージェント向けの本番品質スキルセット。AIアシスタントを実務レベルで動かすための土台。 github.com/addyosmani/agent-… 4. SkillSpector(NVIDIA) AIエージェントスキルの脆弱性・悪意あるパターン・セキュリティリスクを検出するスキャナー。NVIDIAが公開したエージェント安全対策ツール。 github.com/NVIDIA/SkillSpect… 5. codebase-memory-mcp コードベースを永続的な知識グラフとしてインデックス化するMCPサーバー。158言語に対応し、依存ゼロの単一バイナリで動く。 github.com/DeusData/codebase… 6. OpenMontage AIコーディングアシスタントを動画制作の環境に変えるオープンソースシステム。12本のパイプラインと52種のツールを搭載する。 github.com/calesthio/OpenMon… 7. PaddleOCR PDFや画像を、AIが扱いやすい構造化データに変換する軽量OCRツールキット。100言語以上に対応し、LLMへの入力前処理に使える。 github.com/PaddlePaddle/Padd… 8. agentsview Claude Code・Codexなど20以上のコーディングエージェントのセッション履歴・トークン使用量・インサイトを、ローカルで検索・分析できるツール。 github.com/kenn-io/agentsvie… 9. LMCache LLMのKVキャッシュを高速化する専用レイヤー。同じコンテキストの再計算を省き、推論コストと遅延を下げる。 github.com/LMCache/LMCache 10. flue(withastro) 自律エージェントを構築・実行するためのTypeScript製ハーネスフレームワーク。セッション・ツール・スキルに加え、安全な実行のためのサンドボックスを備える。 github.com/withastro/flue
27
354
11
2,093
164,282
My new HackerNoon article reached #6 in top-read stories, with 1.7k+ reads. It shows why AI coding agents need distributed context for microservices — and how archspec provides it before implementation. hackernoon.com/why-sdd-break…
2
10
918
Published Part 2 of my series on why SDD breaks down in microservices. This one is about why I built Archspec: an open-source Claude Code plugin to keep service context explicit with YAML contracts, architecture docs, diagrams, and pre-commit checks. hackernoon.com/why-sdd-break…
62
My article on spec-driven development in microservices made it to the Top 3 on HackerNoon TechBeat by story reads. It's about why LLM coding agents can write good local code, but still miss cross-service rules. hackernoon.com/why-spec-driv… #microservices #ai
3
3
727
Yesterday joined the 📷Podlodka AI Engineers Club meetup. Strong engineers, honest talks about AI in real work: agents, workflows, code review, team practices, and the messy parts behind demos. Great energy in the room. Thanks Podlodka AI and @igrekde for organizing it.
1
3
718
krus210 retweeted
I published Part 1 of a series on why spec-driven development breaks down in microservices. LLMs can write clean code and passing tests for each service, while still missing the rules between services: idempotency, Outbox, batch APIs, state transitions. hackernoon.com/why-spec-driv…
2
1
589
I published Part 1 of a series on why spec-driven development breaks down in microservices. LLMs can write clean code and passing tests for each service, while still missing the rules between services: idempotency, Outbox, batch APIs, state transitions. hackernoon.com/why-spec-driv…
2
1
589
Grateful to have served as an Official Judge & Technical Evaluator for DX-Ray Hackathon 2026 focused on improving developer experience. Reviewed projects across CI/CD, testing, docs, onboarding, dependencies, code review, and environment integrity. Thanks for the opportunity!
16