Gospel A.

Key Skills

Software

Clickworker

AWS SageMaker

Axiom AI

Anno-Mage

Argilla

Top Subject Matter

Multilingual NLP data quality assurance (phonetics, morphology, syntax, transcription/translation, NER).

AI safety evaluation via adversarial prompt design and prompt-response quality assessment.

Multimodal dataset labeling and safety stress-testing for model training.

Top Data Types

Text

Video

Audio

Document

Top Task Types

Entity (NER) Classification

Red Teaming

Fine-tuning

Transcription

Text Generation

Classification

Prompt + Response Writing (SFT)

Freelancer Overview

AI Linguistics & Data Quality Consultant, Veritone Labs Ltd. (Remote). Brings 7+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and N. Education includes Bachelor of Science, Olabisi Onabanjo University (2022). AI-training focus includes data types such as Text, Video, and Audio and labeling workflows including Entity (NER) Classification, Red Teaming, and Fine-tuning.

Labeling Experience

AI Safety & Prompt Evaluation Specialist, Freelance (Remote)

TextRed Teaming

Crafted adversarial and scenario-based prompts to evaluate LLM reasoning quality, bias, and adherence to safety constraints. Simulated dialogue scenarios to test content moderation behaviors and ethical boundary responses. Analyzed model outputs for linguistic accuracy, emotional tone, and safety/compliance violations using structured evaluation templates. • Designed prompt scenarios specifically to elicit unsafe, biased, or non-compliant responses. • Evaluated outputs with QA scoring rubrics covering language and compliance signals. • Developed creative dialogue simulations to test moderation and ethical limits. • Partnered with research teams to refine evaluation templates and improve scoring consistency.

2024 - Present

Speech Model Trainer & Data QA Specialist, Synthetix Research (Remote)

AudioTranscription

Supported ASR and NLU development through targeted speech data annotation and validation. Identified transcription inconsistencies and documented dataset vulnerabilities impacting model training quality. Provided recommended improvement strategies to strengthen transcription reliability. • Performed speech dataset labeling and validation for ASR/NLU workflows. • Flagged and analyzed transcription issues and inconsistencies. • Documented vulnerabilities discovered during data QA. • Suggested improvements to annotation and validation processes.

2025 - 2025

AI Linguistics & Data Quality Consultant, Veritone Labs Ltd. (Remote)

TextEntity (NER) Classification

Provided linguistic QA and data quality for multilingual annotated datasets covering phonetic, morphological, syntactic, transcription, translation, and entity recognition fields. Built linguistic standards, QA guidelines, and documentation to ensure annotation deliverables matched client and technical specifications. Maintained verified 98%+ accuracy through quality reviews and iterative feedback with annotators. • Reviewed annotated transcription, translation, and NER outputs for accuracy and consistency. • Developed annotation training and calibration exercises for phonetic transcription and dialect distinctions. • Automated validation workflows using regex and custom QA scripts. • Produced gold benchmark datasets and error-trend analytics to support evaluation and retraining.

2021 - 2025

Multimodal LLM Trainer, Cognivia Systems (Remote)

VideoText Generation

Annotated and validated large-scale video and text datasets to train multimodal reasoning systems. Designed creative, dialogue-driven test cases to evaluate reasoning accuracy and contextual alignment. Ensured labeled training data supported robust multimodal reasoning behavior. • Performed multimodal annotation and validation across video-text pairs. • Built dialogue-based test inputs to probe reasoning and context adherence. • Focused labeling on improving multimodal reasoning training outcomes. • Supported dataset quality checks to maintain training suitability.

2023 - 2024

Data Annotation & Model Evaluation Analyst, NovaData AI (Remote)

VideoFine-tuning

Supervised multimodal annotation pipelines to ensure structured data integrity and alignment with model training goals. Designed complex and edge-case prompts to stress-test AI safety features within training and evaluation contexts. Produced data insights and quality metrics to support cross-functional improvements for model performance. • Managed multimodal annotation workflow quality and schema compliance. • Built edge-case prompt sets targeting AI safety behavior. • Reported quality metrics and data insights to engineering and product teams. • Ensured labeled data supported model training objectives.

2023 - 2024

Show more

Education

Work History