Project Hedgehog (Handshake AI) - Annotator
Contributed to Handshake AI's Project Hedgehog as an image annotator, producing high-quality object detection labels on a large, diverse image dataset spanning a wide range of subjects, scenes, and visual conditions. Responsibilities included drawing precise bounding boxes around designated target objects, applying the project's labeling taxonomy consistently across edge cases involving occlusion, overlap, low lighting, motion blur, and ambiguous category boundaries, and flagging items that fell outside guideline scope for reviewer adjudication. Worked through a high volume of tasks while maintaining strict adherence to evolving annotation guidelines, calibration updates, and rubric refinements issued throughout the project. Quality measures included alignment with gold standard reference items, inter-annotator agreement checks, multi-pass review cycles, and prompt incorporation of QA feedback to maintain accuracy and consistency at scale.