Role Scope & Requirements- Commercial Experience: 4+ years of commercial software development, with 3+ years focused on Node.js / TypeScript backend development.
- Frontend & Database Stack: Solid Vue 3 (Composition API, Pinia) experience (or readiness to own frontend pilot deliverables), hands-on MySQL, MongoDB, Redis, and message queues (RabbitMQ or Bull).
- Agentic AI Practice: 1+ year of daily experience applying agentic workflows on real projects (e.g., Claude Code, plan mode, sub-agents, parallel sessions in git worktrees).
- Spec-Driven Engineering: Practical experience with at least one spec-driven framework (Spec Kit, OpenSpec, BMAD, Kiro, or a custom operational process).
- Context Architecture & Testing: Ability to write and maintain repository agent contexts (AGENTS.md / CLAUDE.md, skills, rules, hooks, permissions) and perform safe refactoring of legacy code using characterisation tests.
- AI Security & Governance: Clear understanding of AI development risks (handling secrets, personal data protection, code verification).
Nice-to-Have Experience:- Secondary stack exposure: PHP (CodeIgniter 3), NestJS, Fastify, ClickHouse, PostgreSQL.
- Experience with legacy Node.js upgrades (e.g., v12/v14 to current LTS) and spec-driven legacy module migrations with data parity checks.
- Hands-on experience building MCP servers, Claude Code plugins, or GitLab CI pipelines with automated AI review checks.
- Domain exposure: High-load platforms, FinTech, or iGaming (payments, KYC, bonus engines, provider integrations).
Key Responsibilities- Pilot Delivery (AI-Native Mode): Deliver 1–2 pilot project streams from specification to merge request using AI agents, maintaining ownership over architecture, plans, and code quality.
- Spec-Driven Execution: Apply spec-driven workflows (specifications, technical planning, task breakdown, automated verification) with specs versioned in the repository.
- Agent Context & Enforcement: Build repository context structures (AGENTS.md, CLAUDE.md, skills, slash commands, MCP) and convert rules into enforced settings (permissions, hooks, secret isolation, automated linting).
- Verification Setup: Provide agents with deterministic feedback loops (characterisation tests on critical paths, documented local execution commands).
- Contract Specification: Document cross-service contracts (queues, JSON-RPC, shared schemas) in machine-readable formats.
- CI Integration & Knowledge Transfer: Introduce AI-assisted code reviews in GitLab CI, track delivery metrics (cycle times, failure rates), and package learnings into reusable templates to train engineering teams.
Project OverviewThe project focuses on modernising software delivery practices for an established platform. The initiative aims to introduce an AI-native development process - executing pilot projects end-to-end with AI agents and spec-driven development, followed by building shared standards, templates, and an AI SDLC for the wider engineering team.