Question 1

Who is Lukman Nuriakhmetov?

Accepted Answer

Lukman Nuriakhmetov is an engineering and product leader with 16 years of software experience, 8 of them in engineering leadership, across seven industries and six countries. He currently leads engineering on a confidential cross-border payments program. He builds open-source tools for AI-era engineering — Approva, Rhodd, and Codencer — and writes about teams, systems, AI adoption in engineering, and the economics of engineering decisions.

Question 2

What does Lukman build?

Accepted Answer

Three open-source projects for AI-era engineering. Approva is approval infrastructure for AI agent actions (pauses risky agent operations, verifies humans via passkey, issues scoped capabilities). Codencer is an orchestration bridge between planning agents and coding agents. Rhodd compiles architecture definitions in YAML into production-ready Go or Node.js services, Kubernetes configuration, and CI/CD pipelines.

Question 3

What does Lukman write about?

Accepted Answer

Engineering leadership, AI adoption in engineering organizations, senior hiring economics, team structure, career growth in software engineering, and the operating disciplines that make humans and AI work well together. Recent flagship essays cover why organizations that fail at managing people fail at managing AI for the same structural reasons, and the difference between rigorous senior hiring loops and badly bounded evaluation systems that mistake friction for rigor.

Question 4

Where is Lukman based?

Accepted Answer

Between Nice, France and Luxembourg. Holds Passeport Talent (France) and Blue Card (Luxembourg).

Question 5

What is Lukman's engineering background?

Accepted Answer

Sixteen years in software across adtech (Moscow, Paris), chat-commerce (Tokyo), cloud infrastructure (Hannover), geospatial SaaS (Paris), mobile AI commerce (Nice), and cross-border payments (Luxembourg). Has used LLM-based coding agents as part of daily engineering practice since early 2023. Has been through 350+ interview processes as a candidate and hired or participated in hiring 150+ engineers and engineering leaders.

ИИ не меняет сути эффективного управления разработкой — он просто делает ошибки руководства более дорогими

Короткие посты последних недель.

One task in my workflow can be researched in ChatGPT, implemented by Codex, challenged by Claude, and updated through MCP.

I can run several agents in parallel — that does not mean I should let all of them think before I do.

A conversation that follows you across devices is a feature — authority that follows it is a security decision.

Эссе, которые стоят времени.

Bypassable by design: the architecture problem behind AI governance theatre

Инфраструктура, которую я строю.

Вне работы — музыка, менторинг, книги, эссе.