Remote OpenClaw Agent Architect — AI Orchestration

Outlier; Belfast City District; Remote; GBP 43,000 - 61,000; Full time

Benefits offered by this job

Weekly payments

Remote work

Free AI access

Job summary

Outlier invites engineers to shape the future of autonomous AI agents through OpenClaw projects. This freelance, remote role pays up to $50 USD per hour, with weekly payments and flexible scheduling.

Strong candidates bring 2+ years of backend, AI automation, and complex systems experience, plus hands-on API integrations and live-environment work. You will write evaluation rubrics, debug agent traces, and stress test agents while contributing to real-world, multi-turn system tasks across diverse

Qualifications

• 2+ years of backend engineering, AI automation, or complex systems integration.
• Proven ability to build and maintain production-grade software with modular separation.
• Experience in live, non-mocked environments and multi-turn system interactions.
• Excellent attention to detail and ability to provide high-density technical feedback.
• Nice to have: multi-stage coordination tasks with data acquisition to output.
• Hands-on experience integrating agents with live tools like Supabase and Gmail.
• Experience identifying privacy leaks, authority escalation, or prompt injection risks.

Responsibilities

• Write Eval Rubrics
• Debug Agent Traces
• Stress Test Agent

Skills

Backend engineering

AI automation

Complex systems integration

Live environments experience

Technical feedback

Multi-stage coordination

API integration

Memory/state management

Privacy/security awareness

Tools

Supabase

Gmail

APIs

Job description

Outlier invites engineers to shape the future of autonomous AI agents through OpenClaw projects. This freelance, remote role pays up to $50 USD per hour, with weekly payments and flexible scheduling.

Strong candidates bring 2+ years of backend, AI automation, and complex systems experience, plus hands-on API integrations and live-environment work. You will write evaluation rubrics, debug agent traces, and stress test agents while contributing to real-world, multi-turn system tasks across diverse