Emaidiong E.

Key Skills

Software

Internal/Proprietary Tooling

Label Studio

Appen

Other

Don't disclose

Top Subject Matter

Data Annotation

Function Call

Model Evaluation

Top Data Types

Text

Audio

Image

Top Task Types

Transcription

Classification

Bounding Box

Object Detection

RLHF

Evaluation/Rating

Data Collection

Function Calling

Segmentation

Freelancer Overview

I am an AI Data Evaluator and Data Annotator with hands-on experience across LLM response evaluation, search relevance, multilingual text labeling, audio annotation, egocentric video annotation, AI agent testing, and Google Maps/local data contribution. I have worked on projects involving side-by-side model comparison, instruction-following checks, factuality verification, sentiment and intent labeling, transcription, query-result relevance rating, image-to-page match evaluation, and first-person video annotation for object-action interactions. My work spans text, audio, image, video, conversational data, and location-based contributions, with experience in English and Nigerian Pidgin. What sets me apart is my ability to combine strong guideline-following with human judgment, clear written rationales, and real-world data awareness. As a Google Maps Local Guide Level 5, I have contributed 219 verified submissions with over 137,000 views, helping people make better decisions about real places. Alongside my B.Sc. in Psychology, this gives me a strong foundation in observation, behavior, context, and accuracy. I am especially strong at handling ambiguous cases, spotting edge-case issues, applying rubrics consistently, and producing reliable training data that helps improve AI model quality.

Labeling Experience

Business Analyst - AI Response Evaluator (Contract, Turing Remote)

Don't discloseText

Conducted side-by-side (SxS) evaluations of GHD model outputs by comparing multiple completions and selecting the strongest response using rubric criteria. Scored instruction adherence, completeness, relevance, and factual accuracy, documenting misses, gaps, and strengths for each comparison. Verified claims using targeted web research to separate supported information from assumptions. • Performed rubric-based SxS response selection • Rated instruction-following, completeness, relevance, and factuality • Performed web-backed claim verification • Wrote concise rationales explaining evaluation decisions

2026 - 2026

AI Data Evaluator (Contract, Verita Studio)

OtherText

Evaluated AI agents on real workflows across MS Office and Windows tasks to confirm correct execution and outcomes. Wrote and refined prompts for multimodal PC testing and documented results clearly for model validation. Tested websites, software, and products to assess task flow, logic, and user-facing behavior during AI-assisted execution. • Verified AI agent workflow execution correctness • Authored/refined multimodal prompts for testing • Documented test results for validation • Assessed website/software/product behavior against expectations

2025 - 2026

Data Annotator – Ego-centric Annotations (Atlas Capture)

Text

Performed segment-by-segment audits for ego-centric video dataset labels used to train computer vision models in human physical task recognition. Validated object interaction tags and precise timestamps to ensure high-quality temporal alignment with visual actions. Corrected descriptive metadata to achieve 100% alignment between annotations and observed ego-view actions. • Reviewed labels across segments for consistency and guideline compliance • Ensured timestamp and interaction tag accuracy for temporal data • Corrected descriptive metadata discrepancies • Supported dataset quality for CV model training

2025 - 2026

Search Relevance Evaluator - UHRS (Toloka)

TolokaImageClassification

Evaluated image-to-website match tasks by judging whether webpages are relevant to given search queries using platform rubrics. Assessed visual alignment and matching attributes including branding and address details. Checked site functionality and matching quality to support accurate relevance decisions. • Determined image-to-webpage match correctness using rubrics • Verified visual alignment and key matching attributes • Checked webpage functionality for relevance • Ensured consistent relevance scoring across tasks

2025 - 2025

AI Data Evaluator | Data Annotator (Rater-X)

TextClassification

Annotated multilingual conversational datasets (English and Nigerian Pidgin) with sentiment, intent, and conversation-type labels. Scored query–result pairs for relevance based on context, user intent alignment, and usefulness. Created a label schema to better capture code-switching and cultural nuance in conversational data. • Labeled sentiment/intent/conversation-type for multilingual conversations • Evaluated query–result relevance using user intent and usefulness • Extended/structured a label schema for cultural nuance • Applied consistent labeling decisions to dataset examples

2024 - 2025

Show more

Education

Work History