Role Overview
Join our team as a Full-Stack Developer where you'll play a pivotal role in building the Crab Crew platform. This innovative multi-agent system coordinates AI agents, manages client data, integrates with external tools, and delivers tangible outcomes. You'll be working on a greenfield project, contributing to a system that is architected but not yet built.
Responsibilities
- Develop the Agent Framework by building the runtime for defining, configuring, and executing AI agents, including the orchestration layer for task routing and quality assurance.
- Design and build the Client Context Store to provide agents with a comprehensive understanding of each client's brand, goals, tech stack, geography, and history.
- Implement the Shared Task Board using Supabase real-time subscriptions for agent coordination, task claiming, output posting, and dependency tracking.
- Create the Integration Layer to connect agents with external tools and services through MCP tools, APIs, and n8n workflows, including tool definitions and webhook contracts.
- Develop the Knowledge Base by implementing an embedding pipeline for semantic search using Obsidian markdown files and pgvector.
- Establish the Control Plane on Vercel for system monitoring, trigger evaluation, and visibility into the agent platform.
Required Skills
- Proficient in TypeScript/Node.js with experience in strict mode, ES modules, async/await, and building production systems.
- Strong knowledge of PostgreSQL, including schema design, migrations, Row-Level Security policies, and Supabase's PostgreSQL layer.
- Expertise in API Design & Integration, including REST APIs, webhook patterns, and third-party service connections with authentication patterns.
- Experience with Docker for containerized development and debugging.
- Proficient in Git and collaborative development practices.
Nice to Have
- Experience with AI/LLM Integration, including working with LLM APIs and understanding prompt engineering and token management.
- Familiarity with Supabase or Firebase, particularly with real-time subscriptions and auth systems.
- Knowledge of Vector Databases/Embeddings like pgvector or Pinecone.
- Experience with n8n or Workflow Automation tools and self-hosting.
- Comfort with deploying and managing applications on AWS EC2.
- Understanding of MCP (Model Context Protocol) and multi-agent systems.
- Familiarity with Vercel/Serverless architecture and marketing technology tools.