AI Training Specialist (Prompt Engineering & Agent Evaluation) — Confidential | World-Leading Frontier AI Lab (Contract)
Other
Worked as an AI Training Specialist at a frontier AI lab, focusing on prompt-driven autonomous agent workflows. Evaluated agent behavior by executing and recording end-to-end UI interaction trajectories across multiple heterogeneous client applications. Refined prompt structures and validation strategies to improve reasoning quality, execution accuracy, and real-world task reliability. • Engineered multi-step system prompts for autonomous agents • Adapted prompt design and validation for differing UI architectures • Recorded end-to-end UI interaction flows to create reference trajectories • Compared execution paths and improved agent performance via iterative Q&A
2026 - Present