REMOTE RU | Postroy RF
About the project
This is a product SPA that is already in production. The codebase is large: its own UI library of 80+ components, multi-step wizards, real-time chat, maps, and geolocation. The backend lives separately (REST API and WebSocket service), and the frontend is in its own repository.
We are building our work around AI agents. The repository contains CLAUDE.md, detailed documentation in docs/, and project skills for Claude Code: for routing, Redux, react-query, forms, styles, and UI kit. We expect you to work with the agent, not next to it.
Stack
- React 18 + TypeScript (strict, noUnusedLocals/Parameters), build on Vite (SWC)
- Redux Toolkit + redux-persist: state is split into feature slices, there is a compatibility layer with the old structure
- react-query v3 for server data (migration to TanStack v5 is planned)
- react-hook-form for simple forms; multi-step wizard data is stored in Redux and restored from sessionStorage
- CSS Modules + SCSS and design tokens on CSS variables
- React Router, pages are loaded via lazy-loading
- axios: requests go through the sendAsync wrapper with automatic token refresh
- WebSocket: real-time chat (ChatV2)
- Yandex.Maps (ymaps3) and @turf/turf for geolocation and working with zones
- Sentry, authorization via VK ID
- ESLint with --max-warnings 0, Prettier
Responsibilities
- Develop new product features: pages, order and application wizards, personal accounts, filters, catalogs
- Develop the chat (ChatV2) and other real-time scenarios
- Expand the UI kit and design system, layout designs from Figma
- Connect new API endpoints: services, query hooks, cache management
- Gradually refactor legacy code without major rewrites: migration to TanStack Query, moving away from outdated layers
- Support the AI infrastructure of the repository: CLAUDE.md, documentation in docs/, agent skills
Mandatory Requirements
Experience with AI agents: Claude Code and/or OpenAI Codex (mandatory)
- You use the agent daily in real work, not just tried it a couple of times
- You know how to set tasks for the agent: provide context, constraints, and readiness criteria, break down a large task into steps, work in planning mode
- You critically review the results: read the diff, notice hallucinations and unnecessary code, do not send to review what you yourself have not understood
- You understand how CLAUDE.md/AGENTS.md, skills, custom commands, hooks, and MCP servers (Figma, browser) work, and you know how to configure them
- You know when the agent speeds up work and when it's faster to do it yourself
Frontend
- 3+ years of commercial development experience with React
- Confident TypeScript in strict mode
- Redux Toolkit: slices, async thunks, state normalization
- Experience with react-query or TanStack Query: cache, invalidation, mutations
- Forms and validation (react-hook-form), multi-step scenarios
- Layout using SCSS/CSS Modules, adaptive design, pixel-perfect according to Figma mockups
- Working with REST API, understanding JWT authorization (access and refresh tokens)
- Git, working via merge requests, code reviews
Bonus Points
- WebSocket and real-time interfaces (chats, notifications)
- Yandex.Maps or other mapping APIs, geodata
- Library migrations in a large codebase
- Sentry, error analysis in production
- Mini-apps (Max, Telegram)
- Writing your own skills, MCP servers, or automations for AI agents
- Experience configuring tests (Vitest, Playwright)
