At Outlier, I tested ChatGPT, Muse, and other AI agents on everyday tasks, scoring accuracy, helpfulness, and instruction-following.
I recorded agent sessions and wrote reproducible bug reports that captured expected versus actual behavior, severity, and suggested fixes.
As a Quality Assurance / Data Annotation Specialist, I review and label data against strict rubrics, flag unclear instructions, and meet weekly delivery targets.
On my AI Agent Task Benchmark project, I built a 30-task checklist to compare ChatGPT and Muse. I also kept a Prompt Failure Log to sort agent errors by type and severity.
