PETER NGANGA KAMAU

Reviewed and refined AI-generated text and image classifications for a high-volume machine learning training pipeline. Applied annotation rubrics and contextual guidelines across thousands of tasks, achieving consistent high approval rates. Identified recurring biases in motion/spatial descriptions and provided corrective feedback to improve prompts and model behavior. Flagged ambiguous, corrupted, or low-quality data points for escalation, reducing downstream rework, while maintaining steady daily output and style consistency.

Operated Vicon motion capture systems to collect high-frequency kinematic data (up to 240 Hz) for gait, jump, and reach tasks. Produced step-by-step present-tense movement descriptions for research datasets, including trajectories, joint rotations, and relative segment distances from 3D skeletal outputs. Determined exact direction, travel distance, and magnitude of subject movement for accurate spatial documentation. Collaborated with a small team to calibrate equipment, define marker sets, validate capture accuracy, and document protocols and observations in detailed lab reports.