Thaddeus Arama

I review and annotate general, procedural, and first-person video footage for micro1, OpenTrain AI, CloudFactory, and freelance clients. I perform frame-by-frame labeling, object boundary tracking, timestamp review, hand-pose and action annotation, and structured QA against detailed rubrics. I also evaluate RLHF tasks through comparative rankings and written justifications. I use error taxonomies and edge-case classification to identify ambiguities and support accurate AI-training data pipelines. I scored 8.9/10 on a robotic/procedural video annotation assessment. I also built Recodi, a full-stack dashboard tool that automates local operational workflows. My BSc in Economics and Finance and multilingual communication in English, Swahili, Kikuyu, and Gusii support my structured, independent approach to annotation work.