Agentic AI Engineer

AI Engineer – Agentic AI & Frontier Models (Claude Code / Codex / MCP)

Location: San Francisco, CA (Hybrid)

About the Role:

• We are looking for a hands-on AI Engineer who is deeply familiar with today's frontier AI models and the agentic tooling ecosystem built around them.
• The ideal candidate works fluently with Claude Code and Codex on a daily basis and understands how to extend these tools through skills, agents, hooks, and MCP (Model Context Protocol) servers.
• This role combines practical agentic AI development with solid backend engineering fundamentals.

Key Responsibilities:

• Build, configure, and maintain agentic workflows using Claude Code and Codex, including creating skills, agents, and hooks
• Design and develop new MCP (Model Context Protocol) servers, and modify or extend existing MCPs to integrate tools and data sources into agentic environments
• Integrate frontier models (Anthropic Claude, OpenAI models) into applications via API calls, with well-structured prompting strategies
• Develop and maintain backend services, primarily using FastAPI, to support AI-powered applications
• Apply prompt engineering best practices to optimize model outputs for production use cases
• Stay current with the rapidly evolving frontier model and agentic tooling landscape

Required Skills & Experience:

• Frontier model fluency: Strong working knowledge of leading models from Anthropic (Claude) and OpenAI, and hands-on experience using them through tools like Claude Code and Codex
• Agentic tooling expertise: Demonstrated experience creating skills, agents, and hooks within Claude Code / Codex environments
• MCP development: Understanding of how MCPs work inside agentic tools, with the ability to create new MCP servers and modify existing ones
• Backend development: Experience building backend services with Python and FastAPI
• LLM integration: Experience making API calls to frontier models and implementing effective prompting techniques

Nice to Have:

• High-level conceptual understanding of model fine-tuning approaches (deep fine-tuning expertise is not required)
• Experience integrating multiple tools and data sources into agentic loops and automated workflows

What Makes You a Great Fit:

• You are comfortable living inside modern AI developer tooling — not just using Claude or Codex as a chat assistant, but extending and customizing them. You understand the full agentic stack: how skills are authored, how hooks trigger automation, how MCPs expose tools to models, and how to wire it all together with clean backend code.