w
wembrix

Tiffany B.

@wembrix

AI Data Annotation, LLM Response Evaluation Specialist

Nigeria
English, French, Spanish
About me
Data Annotation & AI Response Evaluation Specialist. I help AI teams, researchers, and developers build reliable models with clean, guideline-compliant datasets. Core Expertise: • CV Annotation: 2D Bounding Boxes, Polygons, Segmentation in CVAT • LLM Evaluation: Factuality, Safety, Relevance & Instruction-Following • Quality Control: Error taxonomy, multi-pass validation, strict guidelines Tools: CVAT, Excel, Roboflow | Outputs: YOLO, COCO, JSON, CSV Dedicated to 100% human-in-the-loop precision. Message me to review your guidelines or run a pilot test!... Read more

Skills

w
wembrix
Tiffany B.
Offline • 
Average response time: 1 hour

See my services

Custom Websites
I will bubble io bubble developer bubble website bubble mvp bubble app bubble saas
Data Labeling & Annotation
I will do image annotation, segmentation, bounding box, text labelling and cvat for ml

Portfolio

Work experience

Upwork

Dataset Quality Assurance & Guideline Development

Upwork • Freelance

Jan 2026 - May 20264 mos

Authored comprehensive Standard Operating Procedures (SOPs) and visual edge-case labeling guidelines. • Conducted dataset quality audits, calculating error rates and label consistency metrics. • Remediated noisy annotations, eliminating label drift to deliver verified, training-ready ground truth. Tools: CVAT, Roboflow, Microsoft Excel, Google Sheets.

Fiverr

Computer Vision & Segmentation in CVAT

Fiverr • Freelance

Jan 2025 - May 20261 yr 4 mos

Performed pixel-accurate 2D bounding boxes and multi-vertex polygon segmentation in CVAT. • Managed occlusion, truncation, and class attribute tagging for multi-class detection datasets. • Exported and validated datasets in YOLOv8 (.txt), COCO (.json), and Pascal VOC (.xml) formats. Tools: CVAT, Roboflow, Microsoft Excel, Google Sheets.

Upwork

Healthcare AI Safety & Clinical Guideline Audit

Upwork • Freelance

Sep 2025 - Dec 20253 mos

Evaluated synthetic clinical chatbot responses against patient communication and clinical safety guidelines. • Audited high-risk medical prompts (contraindications, emergency triage, dosage inquiries). • Classified safety risks across a 3-tier severity matrix (Critical, Moderate, Minor) with audit logging. Tools: CVAT, Roboflow, Microsoft Excel, Google Sheets.