Elisha Jackson

LLM Evaluation & AI Quality Specialist with experience supporting AI training, alignment, and evaluation pipelines across multimodal datasets including text, image, audio, video, and documents. Experienced in RLHF-style comparative evaluation, supervised fine-tuning support workflows, instruction adherence assessment, content safety alignment and transcription quality evaluation.

LLM EVALUATION & AI QUALITY SPECIALIST at Outlier AI (2023 – Present)

Subcontract

• Performed RLHF-style comparative evaluation and preference ranking across 500+ AI-generated responses weekly using structured quality rubrics, maintaining 94%+ inter-rater agreement
• Designed and stress-tested 100 prompts to evaluate reasoning quality, factual consistency, instruction adherence, creativity, and alignment behaviour across multiple domains
• Documented recurring edge cases and failure patterns, contributing to updates in evaluation standards and QA frameworks

LLM EVALUATION & AI QUALITY SPECIALIST at iMerit Scholars

Contract

• Evaluated visual outputs for instruction adherence, realism, object consistency, lighting accuracy, and artifact detection across 200 images per batch
• Applied structured annotation frameworks including bounding boxes and markup tools to improve computer vision dataset precision

LLM EVALUATION & AI QUALITY SPECIALIST at Atlas Capture

Contract

• Applied dense and coarse annotation methodologies to support supervised fine-tuning and downstream video-text model training workflows
• Evaluated AI-generated descriptions for ambiguity resolution, grammatical consistency, verb tense alignment, and linguistic accuracy
• Maintained annotation consistency across evolving project guidelines, reducing QA rejection and rework rates by 13%

LLM EVALUATION & AI QUALITY SPECIALIST at TELUS International AI

Contract

• Evaluated and classified 1,000+ content items weekly for safety alignment and policy compliance across multimodal datasets
• Applied moderation and evaluation frameworks to identify misinformation, bullying, unsafe content, and harmful behavioural patterns
• Conducted hallucination review and factual consistency analysis on AI-generated outputs to improve dataset trustworthiness

LLM EVALUATION & AI QUALITY SPECIALIST at Toloka

Contract

• Evaluated and scored 2,000–3,000 images weekly across quality dimensions including sharpness, exposure, noise, artifact presence, and color consistency meeting 100% of QA deadlines
• Applied standardized scoring systems to 10,000+ images, improving dataset uniformity and reducing variance in model training inputs
• Identified and filtered 300+ defective or low-quality images weekly to prevent degraded data from impacting downstream model performance

Bachelor of Science (B.Sc.) in Computer Science – University of Uyo