Bassey E.

Key Skills

Software

Data Annotation Tech

Mercor

Telus

Top Subject Matter

Artificial Intelligence & Machine Learning Operations (MLOps): Expert-level contribution to Reinforcement Learning from Human Feedback (RLHF), prompt engineering, data annotation, and adversarial model testing

Software Development & Technology: Practical experience with frontend web development logic, syntax structure, and digital interface debugging.

Media Strategy & Business Operations: Management of multi-channel media campaigns, strategic workflow optimization, and high-level corporate data reporting.

Top Data Types

Image

Text

Video

Top Task Types

Evaluation/Rating

Transcription

Prompt + Response Writing (SFT)

Data Collection

RLHF

Question Answering

Text Summarization

Object Detection

Text Generation

Segmentation

Entity (NER) Classification

Classification

Freelancer Overview

Since 2023, I’ve had my eye on the AI space and have been actively building my knowledge around prompt mechanics and model behaviour, but I officially transitioned into full-time hands-on work at the end of 2025. Since then, I’ve been working directly on frontier AI models through platforms like Outlier.ai and Handshake.ai. My day-to-day involves running complex multi-turn prompt scenarios and stress-testing models with adversarial inputs to see where they break or hallucinate. A huge part of my job relies on Reinforcement Learning from Human Feedback (RLHF)—meaning I evaluate competing responses, rank them across strict quality dimensions like factual accuracy and constraint compliance, and write clear, logical justifications explaining my scores so AI researchers can effectively tune the models. What really sets me apart is that my timeline in the AI space is backed by a solid technical background in frontend development and operations management. Because I already understand coding syntax and logical workflows, I can evaluate code outputs and complex digital artefacts with a level of precision that non-technical annotators easily miss. I don't just notice if an output feels off; I can pinpoint why the underlying logic failed. This technical eye, paired with my experience managing fast-paced remote workflows, allows me to deliver incredibly clean, highly reliable training data while adapting instantly to shifting project rubrics.

Education

Work History