Data Quality Analyst

About the role

Figure builds autonomous humanoid robots designed to operate with human-level intelligence across home and commercial settings. This role is centered on the AI training pipeline, where you will be responsible for annotating and labeling visual data captured by the company's robots in diverse environments. You will ensure that image and video assets are prepared with high accuracy to support downstream model training and validation. The position requires a methodical mindset to maintain consistency across large and varied datasets. You will work closely with engineering teams to translate labeling requirements into clear, actionable annotations. The schedule for this role is fixed between 1:30 PM and 10 PM, aligning with operational data capture cycles. Success in this position means contributing directly to the quality and reliability of the perception systems used by Figure's robots.

Key facts

Location: USA

Engagement: Full-time, six-month fixed-term agreement with potential for extension

Compensation: Base pay is $30 per hour

Total compensation may include additional components depending on the role

Final offer may vary based on relevant knowledge, skills, and experience

Strong performers may have opportunities to continue beyond the initial term

What you'll do

Apply internal annotation tools to tag objects, body positions, and interactions within images and video captured by humanoid robots.

Partner with machine learning engineers to sharpen labeling standards and streamline tool workflows for efficiency and clarity.

Flag unclear or unusual cases so the ML team can review and resolve them before they affect training outcomes.

Sustain high output while keeping quality consistent across large sets of labeled data under tight production timelines.

Share observations that could improve processes or tool functionality, including suggestions for interface adjustments or automation.

Collaborate with cross-functional stakeholders to understand evolving priorities and adjust annotation focus accordingly.

Monitor data pipelines to identify patterns in annotation issues and work with engineers to reduce recurring problems.

Support the creation of documentation that explains labeling decisions and edge cases for future reference and team alignment.

Contribute to maintaining a structured and auditable annotation history that supports compliance and traceability.

Act as a bridge between raw sensor data and machine learning readiness by ensuring that labeled datasets meet technical standards.

Requirements

• Sharp eye for detail and ability to follow consistent reasoning across varied situations encountered in robot sensor data.
• Comfortable working in software interfaces and picking up new annotation tools quickly with minimal guidance.
• Patient, quality-driven approach with capacity to handle multiple tasks at once without sacrificing accuracy.
• Clear communication in both writing and conversation to articulate annotation decisions and edge cases to technical partners.
• Dependable, self-directed working style suited to a fast-moving team environment with shifting operational needs.
• Fluent English required for both written instructions and verbal coordination with engineering and operations teams.
• Ability to maintain focus during extended sessions reviewing image and video data while upholding high accuracy.
• Willingness to adhere to established labeling guidelines and revise work when feedback is provided by reviewers.

Skills & tools

• Computer-based annotation and labeling platforms used for structured data tagging in robotics contexts.
• Image and video review workflows that involve frame-by-frame analysis and temporal consistency checks.
• Collaboration with technical teams using issue trackers, communication platforms, and versioned documentation.
• Basic familiarity with data formats used in machine learning pipelines, such as annotations for pose, object detection, and segmentation.
• Experience with quality assurance practices such as spot-checking, inter-annotator agreement, and error classification.
• Understanding of human-robot interaction scenarios that benefit from precise semantic labeling of objects and actions.

Practical notes

• The work schedule is fixed between 1:30 PM and 10 PM, requiring reliable availability during these hours.
• This is a full-time position structured as a six-month fixed-term agreement, with performance reviews informing extension eligibility.
• Employment is based in San Jose, Canada, and candidates must be authorized to work in this location.
• Base pay is set at $30 per hour, with total compensation potentially including additional components based on role requirements.
• Final offer details may vary depending on relevant knowledge, skills, and demonstrated experience during the evaluation process.
• High-performing contributors may be considered for continued involvement beyond the initial term based on team needs and outcomes.
• Candidates should be prepared to engage in continuous feedback loops with engineers and supervisors to refine annotation quality over time.
• Attention to consistency, accuracy, and process improvement will be important factors in performance assessment and potential growth.