Full Stack Developer for Vercel Projects

PLEASE READ EVERYTHING # Senior AI Systems Architect | Multi-Agent LLM Orchestration, Hermes, Claude & Autonomous AI Systems ## Project Overview We are seeking an exceptionally skilled AI Systems Architect and Full-Stack Developer to help design, build, and scale a sophisticated multi-agent AI ecosystem. This is not a basic chatbot development project or a simple API integration. We are looking for someone who understands how to architect **multi-layered LLM orchestration systems**, where multiple specialized AI agents can communicate, delegate tasks, reason through complex workflows, utilize external tools, and execute operations autonomously while maintaining appropriate human oversight. Our goal is to develop an intelligent, modular infrastructure capable of supporting multiple businesses, workflows, and operational environments. ## Required Technical Expertise **1. Advanced LLM & Multi-Agent Orchestration** - Multi-agent architectures and hierarchical agent systems - Supervisor/worker agent structures and intelligent task delegation - Agent-to-agent communication and coordination - Dynamic model selection and LLM routing - Context engineering, memory management, and retrieval systems - Tool calling, structured outputs, and autonomous workflow execution - Human-in-the-loop approvals, guardrails, and failure recovery **2. AI Models & Agent Frameworks** - Hermes / Nous Research ecosystem - Anthropic Claude and Claude Code - OpenAI GPT models and APIs - LangGraph, AutoGen, CrewAI, or comparable orchestration frameworks - Model Context Protocol (MCP) - Multi-model architectures combining different LLM providers - Experience with agent evaluation, observability, and optimization **3. Full-Stack & Backend Engineering** - Python and TypeScript / JavaScript - Node.js, Next.js, and React - Supabase / PostgreSQL - Vector databases and semantic search - REST APIs, webhooks, and third-party integrations - Docker, cloud infrastructure, and VPS deployments - Redis, task queues, and asynchronous processing - Secure authentication, permissions, and API management **4. Workflow Automation & Integrations** - n8n and advanced automation workflows - CRM integrations and external SaaS platforms - Voice AI systems (Vapi, Twilio, Deepgram, ElevenLabs) - Event-driven architectures - Background jobs and scheduled agent operations - Custom integrations with business applications ## What You'll Be Building We envision an architecture where a central orchestration layer manages multiple specialized agents, each responsible for specific functions. Examples include: - **Executive Agents:** Interpret objectives, decompose complex requests, and coordinate other agents. - **Research Agents:** Gather, validate, and synthesize information from multiple sources. - **Development Agents:** Assist with coding, debugging, testing, and deployment workflows. - **Operations Agents:** Automate CRM operations, customer communications, and business processes. - **Analytics Agents:** Monitor performance, identify anomalies, and generate actionable recommendations. - **Quality Assurance Agents:** Validate agent outputs, monitor execution, and enforce system constraints. These agents should be capable of working independently within defined permissions while collaborating toward broader objectives. ## Ideal Candidate We are particularly interested in developers who have already built **production-grade agentic AI systems**. You should be comfortable discussing: - When to use single-agent vs. multi-agent architectures - How to prevent infinite loops and unnecessary agent-to-agent communication - How to manage persistent memory and context across long-running workflows - How to balance model capabilities, latency, and inference costs - How to implement secure tool execution and permission boundaries - How to create resilient systems with logging, tracing, retries, and recovery - How to design scalable, maintainable architectures rather than fragile automation chains Also be available during EST time zone please. At least till 2pm. Thank you!