Stephen I.

Key Skills

Software

Scale AI

Appen

Toloka

Top Subject Matter

Large language model training (RLHF)

response evaluation

prompt writing

Top Data Types

Text

Image

Audio

Top Task Types

RLHF

Polygon

Transcription

Classification

Freelancer Overview

AI Data Annotator & RLHF Specialist — Outlier (Scale AI). Core strengths include Scale AI, Remotasks (Scale AI), and Appen. Education includes Bachelor of Science, University of Benin (2018). AI-training focus includes data types such as Text, Image, and Audio and labeling workflows including RLHF, Polygon, and Transcription.

Labeling Experience

AI Data Annotator & RLHF Specialist — Outlier (Scale AI)

Scale AITextRLHF

Evaluated and ranked AI-generated responses for RLHF-style quality dimensions including instruction following, factual accuracy, completeness, tone, and safety for LLM training pipelines. Performed multi-dimension rating using consistent 1–5 rubrics while avoiding bias (e.g., halo effects) across task batches. Rewrote flawed responses with written justifications to create higher-quality training references and prompt inputs. • Ranked responses across multiple quality dimensions. • Applied 1–5 rubric scoring with quality consistency. • Authored diverse prompts to expand training coverage. • Rewrote and justified improvements to weak model outputs.

2025 - Present

Language Data Contributor — Yoruba & Nigerian Pidgin (Appen)

AppenText

Provided specialist Yoruba and Nigerian Pidgin language annotations to support African-language AI training. Performed translation and transcription work and added sentiment labels for language modeling and evaluation datasets. Helped improve coverage for underrepresented West African languages. • Conducted translation/localization tasks. • Recorded/validated voice samples and phonetic transcriptions. • Added sentiment labels for training data. • Supported guideline-based quality checking of contributor work.

2024 - Present

Microtask Annotator — Toloka (Yandex)

TolokaTextClassification

Executed high-throughput text classification and search relevance rating tasks across concurrent projects. Performed image-to-text matching to support retrieval and multimodal alignment use cases. Maintained premium accuracy performance enabling access to higher-rate task pools. • Classified text for multiple projects. • Rated search relevance. • Performed image-to-text matching. • Maintained accuracy above 94% for premium access.

2024 - 2025

Freelance Data Annotator — Remotasks (Scale AI)

AudioTranscription

Performed English audio transcription with speaker diarisation and timestamp-accurate labeling for customer service and interview datasets. Processed large volumes of audio content while maintaining annotation consistency and timing accuracy. Contributed speech-labeled data to support downstream transcription and dialogue understanding models. • Transcribed spoken English content. • Applied speaker diarisation labels. • Ensured timestamp accuracy. • Processed 200+ hours of audio across datasets.

2024 - 2025

Freelance Data Annotator — Remotasks (Scale AI)

ImagePolygon

Completed high-volume computer vision labeling tasks including bounding boxes, object classification, and polygon segmentation for model training. Maintained consistent daily throughput while meeting quality targets across multiple sessions. Produced annotations aligned to platform guidelines to support CV dataset creation. • Drew bounding boxes for labeled regions. • Labeled objects via classification. • Segmented objects using polygons. • Sustained quality above 90% for access to higher-tier queues.

2024 - 2025

Education

Work History