Quality Specialist - Data Annotation

Augmodo•United States, CA

3d•$45,000 - $65,000•Remote

About The Position

About Augmodo Augmodo is at the forefront of spatial computing, using AI and wearable technology to transform physical retail spaces into dynamic digital environments. Our wearable devices capture real-time 3D mapping and inventory data, empowering retail teams to work smarter while giving businesses unprecedented visibility into their physical operations. Role Overview As a Quality Specialist with Augmodo, you will be a part of a small group of individuals who will collaborate across multiple teams to define our Quality goals. This role is about setting the quality standard, not just doing production work. You’ll be a trusted expert—the person we turn to for tough edge cases and the one who creates the benchmark data others are measured against. You'll work closely with our Computer Vision and Data Operations teams to help improve and and define our quality pipeline.

Requirements

2+ years in data annotation and quality assurance, including reviewing and calibrating other annotators' work.
Experience with Encord or equivalent enterprise annotation platforms (e.g., CVAT, Labelbox, SuperAnnotate).
Comfort with complex 2D/3D annotation (bounding boxes, polygons, masks).
Exceptional ability to spot, track, and adjudicate out-of-stock products and edge cases across large video datasets — and to define a defensible standard for them.
Strong verbal and written skills to document standards clearly, justify quality decisions, and collaborate with a global, cross-functional team.

Responsibilities

Curate and maintain critical benchmark datasets used to certify annotators, score vendors, and measure pipeline drift.
Create and update training materials and certification tests for vendor and annotator onboarding.
Periodically evaluate annotators against benchmark data to identify drift and help manage QA metrics and scoring models for the Data Ops Hub.
Conduct rigorous, double-blind quality audits to ensure a gold standard across computer-vision datasets.
Serve as a high-trust annotator to directly handle production labeling when critical (P0) needs arise.