Agentic Engineer - Claude Code
Builds and orchestrates agent-driven apps with Claude Code and other agents; explicitly rejects "vibe-coders" and demands production-grade, under-the-hood skills.
About the Role
Build production applications by orchestrating Claude Code and other coding agents, designing reusable skills and prompt systems, and managing context, memory, and deployments. Deliver complete projects end-to-end, acting as the final quality gate while enabling agent-driven test-first development.
Job Description
Role
You will take projects from specification to deployed software using Claude Code (and other coding agents) to perform most of the implementation. The role focuses on agent orchestration, skill and prompt system authoring, context and memory management, test-driven workflows with agents, and shipping robust production systems.
Key Responsibilities
- Orchestrate multi-agent workflows and design reusable skills (CLAUDE.md, SKILL.md, system prompts).
- Manage context, sub-agents, memory, token usage, and long-running task state.
- Drive TDD with agents using hooks, sub-agents, and verification patterns; enforce agent guardrails and prevent cheating.
- Set up and maintain CI/CD, deployments, databases, secrets management, and monitoring.
- Review, refactor, and approve agent-produced code; act as the final quality gate before shipping.
- Optimize token/cost usage and choose appropriate models for tasks.
Requirements
- Demonstrable production experience shipping apps built largely with Claude Code (not prototypes or toys).
- Fluency with CLAUDE.md, skills, sub-agents, hooks, MCP servers, plan mode, and slash commands.
- Experience with at least one other coding agent (e.g., Cursor, Codex, Copilot, Amp, Cline) and ability to articulate trade-offs.
- Skill authoring and prompt engineering experience creating reusable, consistent agent behaviors.
- Context engineering and memory strategies (RAG, external memory, file-system memory, structured handoffs).
- Strong Git skills (branches, rebases, conflict resolution) and deployment experience (Next.js to Vercel, backends to Fly/Railway, containers to AWS).
- Postgres-class database fluency (Supabase, Neon, RDS, PlanetScale), schema/migrations, indexes, RLS.
- Security best practices: secrets management, env vars, auth flows (OAuth, JWT), input validation, prompt injection mitigation, scoped tool permissions, OWASP Top 10 awareness.
- Basic networking and infra knowledge (DNS, HTTPS, environment isolation, logging/error tracking, uptime monitoring).
- Ability to read and debug code across files, languages, and stacks; strong code review discipline.
- Experience implementing unit and integration testing and agent-driven TDD (pytest, Jest, Vitest, Playwright or equivalent).
- Honest, proactive communication and ownership mentality.
Compensation & Benefits
- Compensation: $10,000+ per month (paid per project) with uncapped upside for high-volume operators.
- Per-project structure with defined scope and deliverables.
- Fully remote and async-friendly work.
- Real autonomy over agent architecture, skills, and workflow.
- Steady pipeline and repeat work for consistent deliverers.
- Direct access to decision-makers and a tooling budget for model APIs, plugins, and infrastructure.
- Peer network of other experienced agentic operators.
How to Apply
Provide a real project you shipped where AI agents wrote most of the code (live URL preferred, repo if open), a skill/CLAUDE.md/MCP server or hook example, your TDD-with-agents approach, and examples demonstrating operations, security, or database work that prove you understand running systems in production.