OpenClaw Agent Builder (Software Developers) in Cardiff

Outlier helps the world’s most innovative companies improve their AI agents by providing human feedback. We collaborate with leading AI organizations to train Large Language Models (LLMs) to function as proactive, multi-step agents. Our projects focus on teaching these systems how to design, coordinate, and optimize complex, real-world architectural workflows.

About the opportunity: Outlier is looking for skilled software experts to help train generative AI models. This freelance role is fully remote and offers flexible hours—you can contribute whenever it fits your schedule.

You may contribute your expertise by:

• Developing objective, verifiable criteria (rubrics) to evaluate system performance and ensure outputs meet strict functional requirements.
• Reviewing system logs and "trajectories" to refactor code, improve execution paths, and reach a "Golden Path" of perfect reliability.
• Testing systems for vulnerabilities, including improper data exposure, unauthorized access escalations, and edge-case failures.

We’re looking for people with:

• 2+ years of experience in backend engineering, AI automation, or complex systems integration.
• Proven ability to build and maintain production-grade software with modular separation (e.g., distinct services for data parsing, logic processing, and reporting).
• Strong command of at least two major languages (e.g., Python, JavaScript, Go, or Java) and experience working with SQL databases.
• Practical experience building for live, non-mocked environments and handling multi-turn system interactions.
• Outstanding attention to detail and the ability to provide clear, high-density technical feedback on complex system behaviours.

Nice to have:

• Expertise building multi-stage coordination tasks where data acquisition leads to reasoned output.
• Hands-on experience integrating agents with live tools such as Supabase, Gmail, and various APIs to solve real-world problems.
• High level of comfort implementing persistent state and session discovery using MEMORY.md to track agent progress.
• Experience identifying subtle failures like privacy leaks, authority escalation, or indirect prompt injections.

Payment: Project work: Earn up to USD $50 per hour for core project work. Additional incentives: On average, Outlier Contributors earn an additional 7.5% on top of the core project rates through Missions — Outlier's version of surge pricing. The top quartile of contributors boost their earnings by an average of 11%. Rates vary based on expertise, skills assessment, location, project need, and other factors. For example, higher rates may be offered to PhDs. For non-core work, such as during initial project onboarding or project overtime phases, lower rates apply. Additional incentives data is based on payments made in the past six months and is updated quarterly. Certain projects offer incentive payments. Please review the payment terms for each project.

PLEASE NOTE: We collect, retain and use personal data for our professional business purposes, including notifying you of opportunities that may be of interest and sharing with our affiliates. We limit the personal data we collect to that which we believe is appropriate and necessary to manage applicants’ needs, provide our services, and comply with applicable laws. Any information we collect in connection with your application will be treated in accordance with the Outlier Privacy Policy and our internal policies and programs designed to protect personal data.

This is a 1099 contract opportunity on the Outlier.ai platform. Because this is a freelance opportunity, we do not offer internships, sponsorship, or employment. You must be authorized to work in your country of residence. If you are an international student, you may be able to sign up for Outlier if you are on a visa. You should contact your tax and/or immigration advisor with specific questions regarding your circumstances.