CV
Anton Serdyuchenko
Software developer building a personal AI-first hub as a path to AI Software Engineering.
Professional summary
Software developer with a backend engineering background, building a personal AI-first hub as a practical path toward AI Software Engineering. Specification-driven development, modular architecture, production-oriented personal systems, and AI-assisted engineering workflows.
Current focus
Designing and operating the AI-first hub: a single-user modular monolith in Go and TypeScript on a single Yandex Cloud VM, used daily for personal task management. Each change is specified, ADR-backed where it matters, and merged through a human approval gate.
Experience
- Owner / Engineer — anton415-hub (personal project)
Single-user modular monolith in Go and TypeScript. Ports-and-adapters per module, PostgreSQL 16 persistence, Yandex ID authentication, Caddy edge, Docker-Compose-on-VM deploys with golang-migrate. The codebase is shaped for AI-readability: enforced module boundaries, small named files, single source of truth per fact.
Evidence: Repository · ARCHITECTURE.md · AGENTS.md
Selected projects
- AI-first Hub In Progress
A personal modular hub for productivity, finance, automation, and AI-assisted engineering — built as a production-oriented single-user system and as a training ground for AI-first software development.
- Todo Production-like
The active module of the AI-first Hub: a single-user task manager with projects, a tree-structured task list, and an editing sheet — designed as a study in productionizing a 'simple' personal module.
- Finance Frozen
A privacy-sensitive personal finance module focused on calm financial decision-making and calculation reliability. Currently frozen — kept as an honest case study of design under privacy constraints.
- Orchestrator Frozen
A module for visualizing and managing the AI-assisted engineering workflow itself — ChatGPT for analysis, Claude for architecture, Codex for implementation, with approval gates and GitHub as source of truth. Currently frozen.
Technical skills
- Languages: Go, TypeScript, SQL
- Backend: chi, pgxpool, PostgreSQL 16, golang-migrate, slog, ports-and-adapters, modular monolith
- Frontend: React, Vite, Tailwind v4, React Router, Astro
- Infrastructure: Docker, Docker Compose, Caddy, Yandex Cloud, Terraform, GitHub Actions
- AI tooling: ChatGPT, Claude Code, Codex, specification-driven development
AI-assisted engineering workflow
-
ChatGPT scopes the problem and drafts the specification before any code is written.
Evidence: Public-site SPEC · Public-site PLAN
-
Claude Code reviews the architecture, writes ADRs, and implements changes inside the layer the file lives in.
Evidence: ADR 0001 — domain boundary · ADR 0002 — repo placement · ADR 0003 — static content
-
Codex picks up visual polish on the frontend after the routes, content, and architecture are fixed.
Evidence: Codex handoff
-
Anton owns every merge to main; one PR per phase, with an Evidence Link in the description.
Evidence: Phase prompt template
-
Boundaries that matter — inward layer dependencies, public/private isolation, deploy guardrails — are enforced by CI rather than by review discipline.
Evidence: AGENTS.md §1.2 · ci.yml · Caddyfile
-
Every architectural decision is written up before the code lands, then linked from the case study and the blog.
Evidence: First post — building the hub · AI-first Hub case study