Mind Robotics is building intelligent robotics for industrial deployment. We believe the fastest path to broadly capable robots is through clearly defined, high-impact environments. We are starting where the need is most acute and the environment is most exacting: the factory floor.

Role Summary

This role is critical to transforming raw collected data into structured, high-quality, and labeled datasets for training our models. You will own the end-to-end annotation program, including defining guidelines, managing quality standards, operating tooling, and ensuring data integrity as volume scales.

Key Responsibilities

Annotation Program Design & Tooling

• Define and maintain annotation guidelines and sub-task ontology in partnership with the ML team.

• Manage annotation tooling, including access provisioning, queue management, and the export pipeline to MCAP.

• Prioritize the annotation queue to align with the ML team’s data requests.

• Explore and integrate annotation platforms (such as Encord) to support customisation and flexibility during the experimentation phase.

Quality Assurance & Data Fidelity

• Develop and maintain the automated quality classification pipeline (Stage 4), including assessing lighting quality and detecting sync errors, frame drops, and hand visibility.

• Monitor Inter-Annotator Agreement (IAA) weekly and run calibration sessions for annotators who fall below the required threshold.

• Build model-assisted annotation pre-labeling, using a lightweight action segmentation model to pre-populate sub-task boundaries for human correction.

Team Management & Throughput

• Hire, onboard, train, and manage the data annotators.

• Establish and track key performance indicators (KPIs) for the annotation program, including reporting weekly on episodes annotated, queue depth, IAA score distribution, and rejection rate.

• Monitor the output of the annotation engine to ensure data is useful for training, focusing on "Quality over Quantity".

Qualifications

• Background: Specialist with 2+ years of experience managing annotation or data labeling teams, ideally with experience in video annotation.

• Technical Expertise: Familiarity with annotation platforms (e.g., CVAT, Label Studio, or similar) and a strong understanding of computer vision components for video processing.

• Strategic Mindset: Ability to treat data collection as a product, balancing immediate operational needs with long-term quality and scaling goals.

• Experience: Proven track record in defining and maintaining structured annotation guidelines and sub-task ontologies.