remote
AI Platform Engineer (Remote)
AI Platform Engineer (Remote)
As an AI Platform Engineer, you will design and evaluate autonomous AI agents across various domains, providing expert human feedback to leading AI organizations. This role involves debugging agent traces, stress testing agents, assessing software architecture, and delivering high-density technical feedback for LLM training.
About the role
Role Overview
Help design and evaluate autonomous AI agents across multiple LLMs, spanning health, education, daily life, and other real-world domains (all coding work). Shape the future of agentic AI systems by providing expert human feedback to leading AI organisations. Help train Large Language Models (LLMs) for complex, multi-step architectural workflows.
Key Responsibilities
AI Agent Evaluation
- Write evaluation rubrics with objective pass/fail criteria
- Debug agent traces to identify failure patterns
- Stress test agents against edge cases, prompt injection, and tool misuse
Technical Assessment
- Assess production-grade modular software architecture
- Analyse multi-turn system interactions and behaviours
- Provide high-density technical feedback for LLM training
Project Workflow
- Create an account and upload a resume/ID
- Complete the onboarding assessment
- Start earning through flexible task assignments
Qualifications
- 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
Preferred (Nice to Have)
- Experience integrating agents with live tools such as Supabase, Gmail, and other APIs
- Familiarity with persistent state and session-tracking patterns
- Experience in identifying privacy leaks, authority escalation, or indirect prompt injection vulnerabilities