Deborah W.

Key Skills

Freelancer Overview

Deborah Wagoha Greater Manchester, UK | [email protected] | 07780163612 PROFESSIONAL SUMMARY Versatile and adaptable professional with nearly a decade of experience across administration, client account management, marketing coordination, frontline support, operations, and AI training tasks. Experienced in working within fast-paced environments, managing stakeholder communication, coordinating projects, maintaining accurate records, and delivering high-quality support across multiple industries. Strong background in communication, organisation, problem-solving, and client-facing work, with the ability to quickly learn new systems and adapt to changing priorities. Recently expanded into AI training and model evaluation, contributing to prompt evaluation, response review, annotation quality assurance, and safety-focused tasks for large language models. Bringing a strong combination of analytical thinking, operational efficiency, and human-centred communication to generalist and AI-focused roles. AI TRAINING & MODEL EVALUATION • RLHF tasks: prompt evaluation, response ranking, and preference review. • Human-in-the-loop review and annotation quality assurance. • Safety and error detection for AI-generated outputs. • Tracking and analysing quality signals to improve training outcomes. • Strong written communication and analytical thinking skills. • Ability to follow detailed guidelines and maintain consistency across reviews. SKILLS & STRENGTHS • Adaptability and quick learning

Labeling Experience

AI Generalist/AI Training Contributor/Scale Labs/ Outlier

ImageRLHF

Contributed to AI training and RLHF projects involving prompt evaluation, response ranking, annotation review, and safety-focused assessments for large language models. Reviewed AI-generated outputs for quality, clarity, consistency, and policy alignment while following detailed guidelines and scoring rubrics. Identified quality concerns, edge cases, and ambiguity issues to support model improvement and maintain high annotation standards.

2025 - Present