At Mindrift, I contribute to 10+ AI training projects, evaluating AI-generated outputs and agent workflows against task specifications and grading criteria. I document evidence-based findings and identify errors, reasoning gaps, task failures, and guideline deviations.
I also assess image and video content for prompt alignment and annotation compliance, including first-person video capture, annotation, quality assurance, review, and correction workflows. I use Python, JSON, scripting, web research, and programmatic scraping where project requirements call for them.
Previously, I enriched and quality-checked datasets containing millions of records at Invisible Technologies. My work has also included coordinating projects and operational processes at Do Any Task, reviewing financial records and controls at A-Plus Life Assurance Company Limited, and proofreading and transcribing publications at Spirit Light Media & Publications.
