Stanley Junior

Detail-oriented AI Training Specialist and Data Evaluator with hands-on expertise in model alignment, Reinforcement Learning from Human Feedback (RLHF), and high-precision multimodal data annotation. Skilled in evaluating complex model outputs, refining prompt responses, enforcing strict taxonomic rubrics, and identifying subtle edge cases across text, video, and kinetic datasets. Proven track record of maintaining high accuracy (>98%), low error rates, and delivering high-quality training signals for frontier AI models under flexible remote workflows.

AI Evaluation & Model Alignment Specialist (2024-01 – Present)

• Evaluate, critique, and refine multimodal AI outputs (video, text, motion capture) to generate high-quality training signals for AI models.
• Apply strict project guidelines and custom evaluation rubrics to score model accuracy, reasoning logic, and adherence to system prompts.
• Perform granular visual and kinetic analysis on complex data clips, identifying subtle human mechanics, timing errors, and visual anomalies.
• Consistently maintain a >98% accuracy rate and minimal rework while executing high-volume evaluation tasks.

AI Trainer & Data Annotation Specialist at Outlier AI (2024-01 – 2025-12)

• Annotated, labeled, and curated multi-format datasets (video, text, audio) to train computer vision, LLM, and machine learning architectures.
• Reviewed and corrected automated model outputs, rewriting vague or inaccurate responses into precise, context-aware training data.
• Analyzed complex edge cases and collaborated with domain leads to refine house-style rules, improve taxonomy, and enhance dataset consistency.
• Evaluated prompt-response pairs to eliminate model hallucinations and ensure alignment with user intent.

Diploma in Computer Science & IT – Mount Kenya University (2023-08)