About me

Vadim Vikulin

I help backend teams (Java/Spring, Python/FastAPI) adopt a methodology that runs on every PR — instead of sitting as a PDF in Confluence.

Author of Use Case Pattern. I lead a backend cluster of 20+ engineers. 17 years in Java/Kotlin.

→ A 30-minute conversation on Telegram

Other contacts:

Over 17 years in Java I watched the same story repeat across six companies: the team grows, delivery slows down, seniors burn out reviewing style. When AI tools landed on my current project, it became clear that chaos scales faster than the team. That's how Use Case Pattern was born: not yet another PDF standard, but a methodology that runs automatically on every PR.

17

years in Java/Kotlin

20+

engineers in the cluster

4

feature teams + platform

6

companies before the current role

«6× fewer defects after formalizing context boundaries before code generation»

Alexey Tolmachev · Senior BSA, 14+ years · a comment under a post on AI and methodology

What I do

Not «how it should be in an ideal world» — but what actually works day to day in a team of 20+ engineers. It has survived a year and a half in production, three shifts of business direction, and people moving between teams.

Methodology

Use Case Pattern

So a team of 20 engineers writes in one style — without an evangelist architect who becomes a single point of failure. Four maturity levels — you pick one to match the module's complexity.

Documentation

Spec-as-code

The specification ships in the same PR as the implementation. It doesn't go stale or drift into «what it was a year ago» — otherwise the review won't pass.

Automation

Claude AI skills

Your standard stops being a PDF in Confluence — Claude applies it automatically on every PR. 12 ready-made skills in the set, for design and review.

Open source

Reference libraries

Open code on GitHub: usecase-pattern, ddd-building-blocks, hexagonal-architecture. You can take it without hiring the author — no vendor lock-in.

Why talk to me

Worth it if you have a backend team of 5+ people (Java/Kotlin or Python) and at least one of these symptoms sounds familiar:

  • The codebase gets hard to maintain faster than you'd like — every feature costs more than the last.
  • AI tools are in place, but three Claude sessions produce three different solutions to the same task.
  • Senior engineers spend code review on style rather than on domain mistakes.
  • Specs go stale before they're even agreed with the business.

I help in four formats:

Service audit 2–3 weeks · a P0/P1/P2 report UCP implementation 2–6 months · team + AI skills Technical mentor recurring · 1-on-1 sparring Code review as a service async · rules cited

Career

17 years, 8 companies, from ETL developer to cluster lead. Here are three key milestones.

  1. now Head of a backend cluster government IT system · 20+ engineers · 4 feature teams + platform
  2. 3 years BelkaCar — Java Team Lead 8 people · Kotlin/CQRS · time-to-market from 2 weeks to 3–4 days
  3. 1.5 years Deutsche Bank — Senior Developer London · 20 people · BDD · deposits

Before that — 4 years at METRO Cash & Carry (EGAIS, Java Concurrency), 4 years at FORS (DWH/BI for the Federal Customs Service and Alfa-Bank), plus three more roles at OTP Bank, Raiffeisenbank, and Avito.

Recognized yourself? Get in touch.

A 30-minute conversation to discuss your context and the format. No «about the company» slides.

→ Get in touch