We’re hiring an AI Engineer to build and deploy production-grade AI agent systems that replace manual, judgment-heavy workflows.
This is a pure IC role focused on one thing: using Claude and modern agent tooling to automate complex operational processes end-to-end.
You’ll join an established technical team:
- A Tech Lead who owns the core app and architecture
- An Automations Specialist who maintains existing workflows (Make, Zapier, GHL, Airtable)
- You, as the AI specialist, identifying where agent systems outperform traditional automations—and shipping those solutions into production
This is not a prototyping role. You’ll build real systems that run the business.
What You’ll Do
Day to day, you will:
- Design and ship AI agent workflows end-to-end using Claude API, Agent SDK, and MCP connectors
- Build custom connectors for platforms without native integrations
- Audit and upgrade existing workflows (Make, Zapier, GHL, Airtable) by layering AI where it meaningfully improves outcomes (routing, classification, edge cases)
- Write production-grade Python and TypeScript (Git workflows, tests, evals, observability, staging)
- Treat prompts as engineered systems (eval sets, versioning, feedback loops)
- Build lightweight internal tools (Slack handlers, Airtable Interfaces, or Next.js apps on Vercel)
- Identify and propose new AI opportunities, with clear effort estimates and ROI
- Stay current with Claude models, Agent SDK updates, and MCP ecosystem developments
- Coach the Automations Specialist on agent patterns and prompt engineering (no formal management)
- Document thoroughly: READMEs, architecture notes, and runbooks for maintainability
Example Project
You might:
- Replace a manual event registration review workflow (~20 mins per registrant) with a Claude-powered agent that:
- Enriches and verifies applicants (LinkedIn + web data)
- Qualifies and routes decisions automatically
- Generates outcome emails
- Provides a one-click human review interface
- Logs everything in Airtable
- Includes test harnesses, monitoring, and runbooks
You own projects end-to-end: from design → production → long-term maintenance.
Remote Remote (Country) Full-time