Full Stack AI Agent Engineer LLM Agents, Claude Code Codex, API Integrations

We are looking for a hands-on AI Agent Engineer to build, deploy, and maintain production-ready AI agents for real-world business workflows. This is not a prompt-engineering-only role. You will work across the full agent lifecycle: understanding the business use case, designing the agent architecture, implementing skills and integrations, deploying agents to production, monitoring them, and fixing issues when they occur. What You'll Do Build custom AI agents for different business use cases and workflows. Design agent architecture, including persona, memory, skills, tools, and communication channels. Build and improve reusable agent skills and templates. Work with Claude Code, Codex, OpenAI and Anthropic APIs, as well as custom LLM agent loops. Integrate agents with external APIs and services such as CRMs, calendars, banking APIs, messaging platforms, and voice providers such as Vapi and ElevenLabs. Implement short-term and long-term agent memory. Work with vector databases, embeddings, metadata, and hybrid search. Set up and work with logging, watchdogs, monitoring, and alerts. Debug production issues and client-reported bugs. Deploy updates using Git, SSH Improve internal tooling and reusable agent infrastructure. What We're Looking For You should have practical software development experience and, most importantly, hands-on experience building AI agents. We'd like to see experience with at least one of the following: Claude Code or Codex Custom agent loops using OpenAI or Anthropic APIs Agent frameworks/platforms such as OpenClaw Production LLM applications with tools, memory, and external integrations You should also understand: How agent loops and tool execution work Agent architecture and separation of persona, memory, skills, tools, and communication channels Short-term vs. long-term memory Vector databases, embeddings, metadata, and hybrid retrieval How to write structured agent configuration/architecture files such as SOUL, RULES, MEMORY, tone-of-voice, and responsibility definitions Linux/VPS environments and production deployment Git and SSH Experience with Python or Node.js is preferred. Experience with Pinecone, Neo4j or similar vector databases is a strong plus. What Success Looks Like You can take an AI-agent task from requirements to a working production deployment with limited supervision. You are comfortable logging into a VPS, inspecting logs, identifying an issue, fixing it, deploying the update, and reporting the result. You work iteratively and quickly: formulate a hypothesis, test it, and produce a result rather than spending days over-engineering a solution. You actively follow developments in AI models, agent frameworks, and tooling and can identify technologies that are actually useful in production. When Applying Please include: 1–2 AI agents or AI products you've built, preferably projects you took from idea to production. A short explanation of their architecture and how they work. Which models, APIs, frameworks, databases, and tools you used. Your experience with Claude Code and/or Codex. Links to GitHub, demos, case studies, or other examples if available.