We are looking for a Senior Full-Stack Developer (AI-Native) to join the project full-time. This person will own the full stack — backend, frontend, and AI/agent layer — with no splitting between projects. They need to be a thinking partner, not an executor — someone the client’s technical leadership wants to work with directly. Your responsibilities will include: * Own the full stack end-to-end — backend, frontend, and AI/agent layer — architecture decisions, performance work, integration, testing — without hand-holding. * Work directly with the client’s technical leadership as a thinking partner. * Build and maintain LLM-powered agent systems with observability, evaluation, and structured tracing. * Cover backend, frontend, and infrastructure independently — no “that’s not my area.” * Use AI-assisted development (Claude Code or equivalent) fluently and extensively as a primary tool.
What we expect from you:Core stack — Backend: * Python — FastAPI, SQLAlchemy, Alembic. * PostgreSQL + PostGIS — spatial data modelling, multi-tenant isolation at the database, session, and repository layers. * Redis — async task queues. * Unit & integration tests — Pytest.
Core stack — Frontend: * React + TypeScript — strict typing; comfortable with generics, discriminated unions, and type inference. * Next.js — App Router, Server Components, server-side data fetching, middleware, Server Actions. * Zustand — store design, selectors, SSR hydration concerns. * Unit & integration tests + E2E — Vitest, Cypress or equivalent.
AI / Agent Layer: * LLM-powered agents — SSE, tool-use patterns, prompt engineering. * Observability & evaluation — structured tracing of agent runs and tool calls, usage tracking, analytics.
Engineering Practices: * Static typing — Pyright as CI gate; type hints on every signature, no escape hatches. * Multi-tenancy — tenant isolation at the database, session, and repository layers. * Git discipline — conventional commits, granular atomic history, clean PRs.
Tooling — AI-assisted development (Claude Code): AI is a primary, everyday tool on this project, not an occasional helper. We expect a candidate who has moved well beyond ad-hoc prompting and treats their AI setup as part of their engineering craft. * AI-native development — non-negotiable. Daily, extensive, fluent use of Claude Code (or equivalent) for reading, writing, refactoring, debugging, and navigating a large codebase. Working without AI in the loop is not how this team operates. * Plain prompting is not enough. “Ask ChatGPT a question and paste the answer” is the baseline we expect candidates to be past. We’re looking for someone who engineers their AI workflow: rich structured context, project conventions and reference examples fed to the model, clear scoping, iteration, and rigorous review of generated output. * Skills, Hooks, Subagents — must understand, must be able to build, and already uses in real work. The candidate should be able to clearly explain what each is, and — critically — has already created and uses them day-to-day, not just heard of them: * Skills — reusable, on-demand instruction/knowledge packages that extend the agent’s competence for a given task type (e.g. scaffold a module, generate a PR, run a review) and encode the team’s conventions. * Hooks — deterministic, automated triggers on lifecycle events (pre-commit, before/after edit, on stop, etc.) executed by the harness itself — e.g. auto-running lint/typecheck/tests on changed files, or guarding protected areas. * Subagents — delegating multi-step or parallel work (large refactors, audits, codebase research) to scoped agents, with verification of their output. * MCP awareness — understands the Model Context Protocol and how to connect AI to real context and tools (codebase, Git, issue tracker, browser, observability) rather than working blind. Hands-on MCP usage is a strong plus. * Structured change management — artifact-driven workflow from exploration through implementation to verification. * Critical, not credulous — reviews everything the model produces, never commits code they don’t understand, knows where AI accelerates the work (boilerplate, tests, scaffolding, exploration) and where it does not (complex domain/architecture logic). AI is a force multiplier for their own thinking, never a copy-paste crutch.
Soft skills / Mindset: * Initiative — proactively asks questions, proposes solutions, flags issues without being asked. * Software Engineering Mindset — understands fundamentals, not just framework recipes; can reason about problems from first principles. * Comfortable with the unknown — says “I don’t know, let me find out,” explores and digs in. * Thinks aloud — communicates where they are, what they’re considering, what they’re unsure about. * Fast — moves quickly without sacrificing quality; comfortable with startup pace and shifting priorities. * Collaborative but opinionated — contributes ideas, challenges assumptions respectfully. * English — B2+ — daily verbal and written communication with the client and their team; working proficiency minimum.
What We Offer: * Work at a Top-employer company (according to DOU 2025). * A strong culture built on empathy, trust, openness, and real care for employees. * Competitive compensation with regular reviews. * Paid vacation and sick leave. * Medical insurance. * Personal learning budget and access to top HR tools, platforms, and practices. * Team events and regular team-building activities. * Flexible hybrid work model with an office in central Kyiv.
