Kelly O.

Key Skills

Software

Data Annotation Tech

Appen

Telus

Top Subject Matter

LLM response evaluation

Rlhf Domain Expertise

safety/bias testing

Top Data Types

Text

Audio

Image

Document

Top Task Types

RLHF

Transcription

Freelancer Overview

Freelance AI Content Trainer & Data Annotator (DataAnnotation.tech | Outlier | Alignerr). Brings 8+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Data Annotation Tech, Outlier, and Alignerr. Education includes Bachelor of Science, Jomo Kenyatta University of Agriculture and Technology (JKUAT) (2023) and Kenya Certificate of Secondary Education, Kenya Secondary School (2019). AI-training focus includes data types such as Text and Audio and labeling workflows including RLHF, Evaluation, and Rating.

Labeling Experience

Freelance AI Content Trainer & Data Annotator (DataAnnotation.tech | Outlier | Alignerr)

Data Annotation TechTextRLHF

Performed weekly evaluation and ranking of 500+ LLM-generated responses for instruction-following, factual accuracy, reasoning quality, and tone to support RLHF pipelines for frontier models. Engineered high-difficulty prompts across STEM, humanities, coding, and creative domains while meeting strict quality rubrics and NDA requirements. Conducted red-teaming and adversarial testing to identify safety violations, hallucinations, and bias patterns that informed model fine-tuning decisions. • LLM response evaluation and preference ranking • SFT training data creation including ideal responses and ranked preference pairs • Red teaming/adversarial testing and quality scoring • Asynchronous collaboration with QA leads to update annotation guidelines

2023 - Present

Digital Content Reviewer & Data Entry Operator (CrowdGen - independent freelance projects)

AudioTranscription

Transcribed and time-stamped 150+ hours of audio and video content annually across English, Swahili, and Spanish while keeping word error rates (WER) below 2%. Labeled and categorized 10,000+ image, text, and audio samples across diverse AI training projects including object detection, intent classification, and dialogue act tagging. Reviewed machine-generated content for grammar, factual accuracy, and policy compliance, and mentored new contributors to meet annotation standards. • Verbatim/clean-read transcription with timestamping • Multi-modal labeling across image/text/audio datasets • Grammar, factual accuracy, and policy compliance review • Contributor onboarding and quality/rubric alignment

2021 - 2023

Remote Quality Assurance Specialist & Research Assistant (Appen | Telus International | Prolific)

AppenText

Reviewed and validated 300+ annotated data packets per week including text classification, sentiment labels, and named entity tags to reduce inter-annotator inconsistency. Performed structured web research and fact-checking by verifying claims against authoritative sources and flagging misinformation with supporting evidence. Assessed search engine result relevance and page quality (EEAT) for query-level evaluation while conducting cultural sensitivity and linguistic quality reviews across English, Swahili, and Spanish. • Text classification, sentiment, and NER validation • Fact-checking and source credibility verification • Relevance/page quality assessment (EEAT standards) • Cultural sensitivity and linguistic review across multiple languages

2019 - 2022

Education

Work History