ML Data Operations Engineer - AI-Assisted Annotation

Summary

The Gatos team is a cross-discipline production and research data annotation team within Video Engineering. We deliver the labeled data that algorithm teams use to build and evaluate features for millions of devices and users around the world. We're looking for an operations engineer who will own annotation projects from the first request to delivered data. That includes a growing set of AI-assisted and agent-based workflows, where models do much of the labeling and people guide, review and verify the results.

Minimum Qualifications

BS and a minimum of 3 years relevant industry experience. Experience delivering data for machine learning, such as data annotation, data operations, or technical program management. Experience defining annotation tasks, labeling guidelines, and quality metrics. Experience serving as the accountable owner for projects that span multiple teams, from ambiguous requirements to delivery. Experience working hands-on with data and technical tools, such as writing basic SQL queries or scripts, analyzing data in spreadsheets, or configuring labeling platforms.

Preferred Qualifications

Experience with model-assisted labeling, pseudo-labeling, or agent-based workflows, including judging where people must stay in the loop. Experience with Python or SQL for data analysis and automation. Familiarity with computer vision data such as images, video, depth, or 3D. Experience building reports or dashboards for cross-functional partners. Clear written and verbal communication, including status updates and escalations to senior leaders. Works effectively across time zones with internal teams and external partners. Defines success with customers before starting the work, and adapts as their needs change.

Description

In this role you will: Work with algorithm teams to define what success looks like for each data request, and translate it into a clear task specification, labeling guidelines and a quality bar. Own annotation projects end to end, from pilot through production to delivery, and be accountable for quality, schedule and the usefulness of the data to the customer. Design, implement, and run AI-assisted labeling workflows: models pre-label, automated steps handle routine work, and people focus on judgment and verification. Measure the impact on quality, throughput and cost. Set acceptance criteria, run quality checks and audits, analyze error patterns, and close the loop with guideline owners and labeling teams. Track project status and risks, and raise issues early with a clear ask and a root cause. Collaborate with data scientists, software and infrastructure engineers, data collection teams, business operations, and external labeling teams across time zones. Improve how the team works by building and sharing tools and workflows that use AI.