Computer Vision Data Engineer

This role is FULL-TIME REMOTE from Latin America in the Pacific Time zone.

Hey, we're Glacier! A Series A startup based in San Francisco tackling one of the world's most pressing problems: trash. Did you know that in the US, we send over half of our recyclables to the landfill? We're working to fix that. In doing so, we'll also be reducing carbon emissions, energy consumption, and depletion of natural resources.

Glacier builds custom sorting robots designed to sort apart recyclables as well as AI-powered business analytics that enable recyclers to superpower their plants and improve our society's circularity.

From major CPG companies like Colgate and Amazon to municipal recycling facilities, our clients trust us to turn recycling data into actionable insights. Our technology has been recognized as one of TIME's Best Inventions and featured in a TIME documentary, TechCrunch, Fortune, and CBS.

Your Role:

We're looking for a talented computer vision data engineer to own the data layer that powers our computer vision models. Great models start with great data, and you'll build the datasets, pipelines, and internal tools that let our ML team move fast and train on high-quality, well-curated data. This is a strong fit for someone who loves building software and internal tools as much as working with data.

Your Responsibilities:

• Own our image and video dataset lifecycle: taxonomy, curation, collection workflows, and quality, so our models train on clean, well-structured, error-free data.
• Design and maintain the data pipelines that move, transform, version, and sync our datasets (e.g. DVC, annotation platforms, cloud storage).
• Build internal tools and software that make dataset creation, labeling, and curation faster and more reliable for the ML and labeling teams.
• Coordinate closely with our labeling team to ensure data is well curated, consistently annotated, and error-free.
• Improve how we collect and prioritize training data over time, including model-assisted and active-learning approaches.

Requirements:

• 2+ years of Machine Learning experience
• 2+ years building software applications or internal tools, with strong expertise in Python
• 2+ years of experience working with image or video datasets
• Hands-on experience designing and maintaining data pipelines (ETL/ELT, workflow orchestration, data versioning).
• Strong SQL and experience working with structured data
• English fluency, as you'll be working with a US-based team (B2 or higher).

Nice to Have:

• Computer vision background or experience supporting CV/ML teams
• Web application development (React + Django, Next.js)
• Experience with annotation platforms (Supervisely, CVAT, Label Studio) and data versioning tools (DVC).
• Experience with cloud storage (S3) and building backend services (FastAPI).
• Experience integrating open-source tooling and contributing to internal platforms.
• Experience working with US companies or clients.