Samuel M.

Key Skills

Software

Clickworker

Labelbox

Micro1

Mindrift

OneForma

SuperAnnotate

Telus

Toloka

Top Subject Matter

AI model response evaluation and RLHF-style quality annotation

Data validation, anomaly identification, fact-checking, and quality assessment workflows

Regulatory Compliance & Risk Analysis

Top Data Types

Text

Audio

Document

Top Task Types

RLHF

Object Detection

Transcription

Evaluation/Rating

Data Collection

Freelancer Overview

I have experience working in AI training and evaluation through my role as an Advanced AI Data Trainer at Invisible Technologies, where I participated in Reinforcement Learning from Human Feedback (RLHF) workflows. My responsibilities included performing side-by-side evaluations of AI-generated responses, fact-checking outputs through online research, rating and ranking responses based on quality guidelines, and writing structured justifications to support evaluation decisions. This role strengthened my attention to detail, analytical thinking, and ability to follow strict annotation standards while working in a fast-paced remote environment. In addition to AI evaluation work, I also have a strong research and data-handling background from my experience in healthcare and academic research. As a Research & Teaching Assistant at the University of Ghana, I analyzed datasets using Excel and SPSS, supported research projects, and prepared reports and presentations for academic use. Combined with my healthcare background in Occupational Therapy, this has made me highly comfortable working with structured information, validating data accuracy, and learning specialized terminology quickly.

Labeling Experience

Advanced AI Data Trainer (2-month contract) at Invisible Technologies

TextRLHF

Performed side-by-side evaluations and fact-checking for AI model responses against provided evaluation guidelines. Rated, ranked, and annotated responses to support Reinforcement Learning from Human Feedback workflows and improve model quality and alignment. Wrote consistent justifications to ensure auditability of the evaluation and annotation decisions. • Assessed accuracy, relevance, tone, and safety for each AI response • Verified outputs via online research and browser-based checks to flag inaccuracies • Produced rankings/ratings and supporting rationale for each decision • Used internal platforms and collaboration tools to complete evaluation and QA tasks

2024 - 2024

Education

Work History