Tochukwu Onwukwe

I work at the intersection of AI training, data annotation, and workflow automation, helping build the datasets, evaluations, and systems that make AI models more accurate, safer, and genuinely useful in production. Data Annotation & AI Training Over the past couple of years, I’ve worked across a range of annotation and evaluation platforms such as: micro1, Outlier, Turing, Test IO, Atlas Capture, Innodata, and CVAT; each teaching me a different piece of what “good data” actually means. On micro1, I’ve completed a large volume of task instances, maintaining a consistently high accuracy rating with full approval across peer reviews. My work has spanned computer vision annotation (multi-point rider and animal skeleton labeling for action recognition tasks), structured bug severity assessments, text annotation with detailed labeling rules, and LLM response evaluation and ranking. What I’ve come to appreciate most in this work is precision under ambiguity, most annotation tasks look simple until you hit the edge case that breaks the rule you were given, and figuring out how to resolve that consistently is really the whole job. That attention to detail is also why I’ve been able to maintain strong accuracy scores across platforms with very different guidelines and QA standards.