Why Leaders Return to Code — Nikita Belokopytov

Nikita Belokopytov connects his growth as a leader with learning to understand his own behavior. When his desire to replace a manager met honest feedback, he asked colleagues for their views, sought help from HR, and deliberately changed his communication habits. Technical confidence gave way to a motivation to help people grow. He later learned that the right degree of autonomy and oversight depends on the person and the work.

At Playrix, Nikita owned the technical side of Gardenscapes: optimization specialists needed help removing obstacles, while quality, releases, and CI/CD needed established processes. Introducing Cursor showed that faster code generation delivers different value to engineers with different starting points. An agent-based bug triage project became useful after they introduced a state graph, tools, and an isolated environment. The workflow had to constrain the agents' unpredictability.

Personal experiments with Gemini and Firebase Studio revealed how much product development had changed. After completing a transformation at Playrix, Nikita chose practical AI work and joined Bluefish to reshape its development cycle. Management experience remained useful in an engineering role: finding the right people, aligning on changes, and taking responsibility. He presents fewer management layers as his own prediction for smaller autonomous teams, rather than a universal rule.

At Bluefish, engineers remain accountable for changes to the core system, while client-facing employees create their own tools. An autonomous agent takes small Jira tasks through to a pull request; a person decides whether the work is needed. Another experiment exposed the limits of manual review and expensive correction loops between agents. Autonomy therefore needs spending, time, and iteration limits. The episode closes with engineers' responsibility for product outcomes and managers' need for deep technical understanding accompanied by an appropriate communication style.