Kathy Travis

Results-driven AI Data Annotation and Evaluation Specialist with proven expertise delivering high-accuracy labeled datasets at scale. Demonstrated track record of improving inter-annotator consistency by 25%+, maintaining 98%+ accuracy across high-volume annotation projects, and reducing data rework by identifying systematic labeling errors early. Expert at interpreting complex guidelines for ambiguous edge cases and translating them into actionable quality improvements.

Trained in AI/ML, Generative AI, and Prompt Engineering with hands-on experience in RLHF workflows, LLM response evaluation, and content safety assessment. Trusted to own quality assurance for mission-critical datasets used in production AI systems.

AI Data Annotation & Evaluation Specialist - BELAY - Remote/Virtual

(2023-01)

• Deliver high-fidelity labeled datasets at scale: annotated 200+ images using CVAT with 2D bounding boxes, maintaining 98%+ accuracy across computer vision projects.
• Enforce strict quality standards: reviewed and validated 300+ text and image annotations, identifying and correcting labeling errors that prevented model performance degradation.
• Improved team consistency: developed annotation decision trees for ambiguous cases, reducing inter-annotator disagreement by 25% and accelerating labeling velocity by 20%.
• Supported content safety: evaluated LLM outputs against safety policies, flagging hallucinations and policy violations with 95%+ detection rate.
• Optimized QA workflow: implemented systematic error-detection process that reduced dataset rework by 40% and saved 200+ hours of re-annotation effort.
• Adapted rapidly: mastered task-specific guidelines within 48 hours and consistently met quality benchmarks across 5+ concurrent projects with shifting requirements.

Customer Service & Data Support - Brickman Properties Limited

(2022-06 - 2022-09)

• Resolved 100+ customer disputes with 95%+ satisfaction rate through structured problem-solving and clear communication.
• Maintained 100% accuracy on communication records and data entry across high-volume customer interactions.
• Built foundation for precision work: developed systems for tracking and organizing customer data with zero errors over 4-month tenure.

Records and Data Support Intern - Ministry of Lands, Housing and Urban Development, Thika

(2017-01 - 2018-12)

• Processed and validated 300+ title deed records with 100% accuracy, ensuring government compliance and data integrity.
• Streamlined record-keeping: organized and indexed 500+ documents, reducing data retrieval time by 50%.
• Demonstrated meticulous attention to detail in mission-critical government records environment.

Diploma - Information Technology - AIRDT Shule, Wote (2025-01 - 2025-12)

Diploma - Cosmetology - Nairobi Institute of Business Studies (2021-01 - 2021-12)

Kenya Certificate of Secondary Education - St. Tito Secondary School (2017-01 - 2017-12)