Job description

About Mind: Mind Robotics is building robots that learn from real-world experience, and the data platform is what turns raw experience into training signal. Every day, sensor data flows in from multiple sources — arriving in different formats, at different volumes, and with different quality bars. This team owns what happens next: ingesting that data reliably, validating and curating it for quality and diversity, powering annotation, and building the automated methods that will increasingly do that labeling work. About the role: As a Software Engineer working on the data platform, you'll be one of the founding engineers on the team that owns our application layer, reporting to our Head of Application Engineering. The systems work today; your job is to help harden and productionize them for scale — ingesting from a growing number of sources and field sites, and turning raw captures into training-ready data faster and at higher quality. This is early, hands-on, 0-to-1 engineering: you'll build the pipelines and tooling that determine what data actually makes it into our models, and see the effect on robot behavior. You will: - Build data ingestion pipelines from multiple sources, including our own field capture and teleop stacks - Build automatic data validation systems that catch quality issues before data reaches annotation or training - Build and improve annotation ingestion, tooling, and workflows to increase labelling efficiency and throughput - Own data quality and diversity — build the systems and metrics that tell us what data we have, what's missing, and where to focus collection - Explore and build automated annotation methods, including running computer vision and vision-language models for tasks like depth estimation, hand tracking, and auto-labeling, to reduce manual labelling load over time - Support the platform — debug issues in live pipelines, instrument for observability, and work directly with research and annotation partners who depend on your software Requirements: - 2+ years of software engineering experience building production systems - Strong programming fundamentals and comfort working across the stack — services, data pipelines, and applied ML tooling - Experience with at least one of: real-time or large-scale data pipelines, ML data workflows, computer vision, or data infrastructure - Bias for ownership: you've taken features or systems from prototype to production and supported them in the field - Clear communication and close collaboration with product, research, and annotation/operations partners - Hands-on experience with sensor data (video, depth, IMU, force/torque) and the infrastructure to process, validate, and label it at scale is a plus - Experience with streaming or near-real-time data pipelines is a plus - Familiarity with ML data workflows (datasets, labelling, evaluation) is a plus - Experience running or integrating computer vision or vision-language models (e.g., depth, pose/hand tracking, open-vocabulary detection, auto-labeling) is a plus

Requirements

Must-haves

Build data ingestion pipelines from multiple sources, including our own field capture and teleop stacks
Build automatic data validation systems that catch quality issues before data reaches annotation or training
Build and improve annotation ingestion, tooling, and workflows to increase labelling efficiency and throughput
Own data quality and diversity — build the systems and metrics that tell us what data we have, what's missing, and where to focus collection
Explore and build automated annotation methods, including running computer vision and vision-language models for tasks like depth estimation, hand tracking, and auto-labeling, to reduce manual labelling load over time
Support the platform — debug issues in live pipelines, instrument for observability, and work directly with research and annotation partners who depend on your software
2+ years of software engineering experience building production systems
Strong programming fundamentals and comfort working across the stack — services, data pipelines, and applied ML tooling
Experience with at least one of: real-time or large-scale data pipelines, ML data workflows, computer vision, or data infrastructure
Bias for ownership: you've taken features or systems from prototype to production and supported them in the field
Clear communication and close collaboration with product, research, and annotation/operations partners

Nice-to-haves

Hands-on experience with sensor data (video, depth, IMU, force/torque) and the infrastructure to process, validate, and label it at scale is a plus
Experience with streaming or near-real-time data pipelines is a plus
Familiarity with ML data workflows (datasets, labelling, evaluation) is a plus
Experience running or integrating computer vision or vision-language models (e.g., depth, pose/hand tracking, open-vocabulary detection, auto-labeling) is a plus

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

Platform: ashby
Effort: quick

Benefits

Benefits detected in the listing.

Vision