About this role
Analyzing daily review queues and conducting judgment-heavy quality assessments on annotated data makes up the core of the daily work. Operating within a multidisciplinary team spanning autonomy, computer vision, and machine learning, this practitioner ensures ground truth quality for warehouse intelligence platforms driven by autonomous drones and existing equipment.
Ownership spans the decision rules, verdict taxonomies, and performance trackers that differentiate between annotation drift, model regressions, and genuine field degradation. Someone who thrives on analytical rigor, exercises sound judgment under ambiguity, and communicates clearly in writing will succeed in this environment.
What you will do
- Manage the daily review queue to produce in-house verdicts and root cause analyses for complex data that cannot be outsourced.
- Develop, version, and refine decision rules, quality guidelines, and verdict taxonomies for assigned categories.
- Build and maintain performance trackers to measure error rates over time by facility, site, equipment, and data format against established baselines.
- Investigate flagged anomalies independently and run root cause analysis to distinguish between labeling errors, system bugs, and real-world degradation.
- Query and analyze annotation data directly using Python and SQL to test hypotheses and validate system behavior.
- Translate findings into concrete updates for standard operating procedures, instruction changes, and annotation tool improvements.
What you need
- Hold a Bachelor of Science in Computer Science, Engineering, Electrical Engineering, or possess equivalent practical experience.
- Bring two to five years of professional background in machine learning quality assurance, annotation quality, data quality, or analytics within an artificial intelligence or machine learning organization.
- Demonstrate hands-on experience working with annotated machine learning datasets and evaluating label quality.
- Possess a strong understanding of statistics along with the ability to generate competing hypotheses and isolate evidence.
- Show proficiency in querying and analyzing data independently using Python and SQL.
- Display demonstrated experience finding, diagnosing, and reporting data anomalies to technical audiences.
- Exhibit excellent written communication skills for authoring quality guidelines, decision rules, and labeling taxonomies.
Nice to have
- Two or more years of experience in quality assurance, data quality, or product support.
- One or more years of experience utilizing enterprise-grade ticketing systems such as Jira.
- Computer vision annotation experience as a reviewer or auditor covering bounding boxes, polygon segmentation, or classification.
- Background in warehouse automation, robotics, or computer vision applications.
- Experience with dashboarding or business intelligence tooling for recurring reports.
Skills and tech stack
Questions candidates ask
Is the role remote?
The position is open to remote work based in India.
What tools and languages will I use?
You will query and analyze data directly using Python and SQL, and track work across Jira.
What will I work on first?
The first assignment focuses on the warehouse forklift vision program, where barcode readability and localization analysis are the immediate need.