Candace Fletcher

Key Skills

Software

CVAT

Data Annotation Tech

Dataloop

Labelbox

Label Studio

Mercor

OpenCV AI Kit (OAK)

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Text

Video

Top Task Types

Audio Recording

Classification

Evaluation Rating

Fine Tuning

Prompt Response Writing SFT

Freelancer Overview

I am a seasoned Data Labeling and AI Training Specialist with over eight years of hands-on experience supporting large-scale machine learning and language model development. My expertise spans text, image, audio, and code annotation, where I’ve contributed to projects improving reasoning, factual accuracy, and alignment of modern AI systems. I have worked with leading platforms such as Mindrift, Appen, Scale AI, and Toloka, delivering consistently high-quality datasets for natural language processing, computer vision, and generative AI research. With a strong foundation in analytical thinking, quality assurance, and structured data evaluation, I excel at identifying edge cases, logical inconsistencies, and ambiguous task definitions. My work emphasizes precision, ethical data practices, and context-sensitive interpretation. I’m particularly skilled in refining annotation guidelines, auditing complex datasets, and mentoring junior annotators to maintain consistent standards. Driven by curiosity and rigor, I bring both technical depth and human insight to every AI training and evaluation project I support.

Labeling Experience

Technical Document Tagging and Named Entity Recognition (NER)

Opencv AI Kit OakDocumentClassificationQuestion Answering

Annotated technical engineering and scientific documents for structured entity recognition tasks involving formulas, materials, and system components. Labeled hierarchical relationships between entities and attributes to enable better document understanding and retrieval. Conducted QA reviews for contextual accuracy, overlapping entities, and domain-specific vocabulary consistency. Supported a multilingual dataset pipeline with English and technical translation checks.

Annotated technical engineering and scientific documents for structured entity recognition tasks involving formulas, materials, and system components. Labeled hierarchical relationships between entities and attributes to enable better document understanding and retrieval. Conducted QA reviews for contextual accuracy, overlapping entities, and domain-specific vocabulary consistency. Supported a multilingual dataset pipeline with English and technical translation checks.

2025

Audio Emotion Recognition and Speech Tagging for Conversational AI

MercorAudioTranslation LocalizationData Collection

Performed emotion tagging and acoustic feature annotation for multilingual conversational datasets. Classified tone, pitch, and intent across diverse emotional categories (joy, frustration, neutrality, etc.). Supported voice model fine-tuning by flagging inconsistencies and ambiguous emotional states. Maintained strict adherence to privacy and linguistic sensitivity standards. Annotated over 40,000 clips with 97% consistency verified by QA reviewers.

Performed emotion tagging and acoustic feature annotation for multilingual conversational datasets. Classified tone, pitch, and intent across diverse emotional categories (joy, frustration, neutrality, etc.). Supported voice model fine-tuning by flagging inconsistencies and ambiguous emotional states. Maintained strict adherence to privacy and linguistic sensitivity standards. Annotated over 40,000 clips with 97% consistency verified by QA reviewers.

2025 - 2025

Visual Object Detection and Segmentation for Autonomous Systems

CVATVideoPolygonSegmentation

Led a visual labeling team annotating over 250,000 frames of street-level and drone imagery to train object detection models for vehicle, pedestrian, and infrastructure recognition. Ensured spatial and temporal accuracy in bounding boxes, polygons, and segmentation masks. Conducted QA checks for occlusion, depth, and object continuity, maintaining consistent labeling across frames. Collaborated with engineers to calibrate precision metrics and adapt annotation tools for 3D bounding tasks.

Led a visual labeling team annotating over 250,000 frames of street-level and drone imagery to train object detection models for vehicle, pedestrian, and infrastructure recognition. Ensured spatial and temporal accuracy in bounding boxes, polygons, and segmentation masks. Conducted QA checks for occlusion, depth, and object continuity, maintaining consistent labeling across frames. Collaborated with engineers to calibrate precision metrics and adapt annotation tools for 3D bounding tasks.

2024 - 2024

LLM Prompt Evaluation & Text Classification (Language Model Alignment Project)

LabelboxTextText GenerationRLHF

Served as a senior text annotation specialist for a large-scale LLM fine-tuning project focused on English-language reasoning and response alignment. Evaluated model outputs across various domains including technical writing, dialogue, and creative generation. Tasks included scoring factuality, coherence, tone, and safety adherence. Created gold-standard examples for supervised fine-tuning (SFT) and developed reviewer training materials. Maintained over 98% accuracy in audit checks and contributed to improved model alignment and contextual understanding.

Served as a senior text annotation specialist for a large-scale LLM fine-tuning project focused on English-language reasoning and response alignment. Evaluated model outputs across various domains including technical writing, dialogue, and creative generation. Tasks included scoring factuality, coherence, tone, and safety adherence. Created gold-standard examples for supervised fine-tuning (SFT) and developed reviewer training materials. Maintained over 98% accuracy in audit checks and contributed to improved model alignment and contextual understanding.

2022 - 2023

Education

Work History