I build and maintain production web products end to end: domain model and API, admin panels and storefronts, payments, CI/CD and deployment. Open to freelance and contract work, fully remote.
Most of my commercial code lives in private client repositories, so here is a summary.
Platform with payments and an online store
Kotlin, Spring Boot, MariaDB · Nuxt, Vue 3, TypeScript · Docker, GitHub Actions
- Built the financial core: transactions, gift certificates, refunds with bank status reconciliation, royalty and acquiring-fee calculations — all covered by tests.
- Designed and shipped the Shop module from scratch: domain model, API, admin, storefront, checkout with payment, customer notifications, permissions.
- Integrated acquiring behind an adapter and a feature flag — switching payment providers without touching business logic.
- Set up CI/CD for a monorepo: builds only changed subprojects, test gate before deploy, fast rollback, migration checks.
- Audited the codebase inherited from a previous team, wrote the spec and estimates for two milestones. 700+ commits — over half of the repository history.
Photo platform with face search
Python, FastAPI, PostgreSQL, Redis · React, Next.js, TypeScript · Milvus, InsightFace · GitLab CI
- Developed a five-service platform: BFF, image upload, selfie search over face embeddings, admin and client frontends.
- Ran a security audit of authentication and public APIs, delivered a prioritized report before release.
- Migrated payments to a new acquiring provider, reworked pricing with regional overrides.
Corporate ERP (Galaktika) — about a year
Business logic in VIP, SQL reporting and query optimization, FastReport print forms, automating manual accounting processes together with business stakeholders.
- Backend: Python (FastAPI, Django/DRF, SQLAlchemy), Kotlin (Spring Boot)
- Frontend: React / Next.js, Vue / Nuxt, TypeScript, Tailwind CSS
- Data: PostgreSQL, MariaDB / MySQL, Redis
- Infra: Docker, GitHub Actions, GitLab CI
- Also: Telegram bots (aiogram), LLM integrations (OpenAI / OpenRouter)
Before development I spent 8 years in rail logistics, up to head of a logistics department — so I'm used to talking with business stakeholders and turning their processes into working software.
- Small, reviewable milestones with a working demo at each step
- Tests and Docker from day one, so the project runs the same everywhere
- Human-in-the-loop for anything AI writes into real data

