Key Responsibilities

Model Evaluation & Ground Truth Systems

• Own the evaluation definitions, metrics, and ground truth standards for computer vision systems operating on satellite imagery.
• Design and maintain benchmark datasets and evaluation protocols for tasks such as:
• Object detection
• Change detection
• Segmentation
• Geospatial feature extraction
• Ensure evaluation frameworks reflect real-world geospatial use cases and edge cases.

Evaluation Infrastructure

• Build and maintain evaluation harnesses that automatically run experiments across models and datasets.
• Publish experiment results to a centralized system (e.g., Weights & Biases or similar experiment tracking tools).
• Develop reproducible pipelines that enable researchers to compare model performance across versions and datasets.

Internal Tooling & Dashboards

• Develop lightweight internal tools and dashboards that allow engineers and researchers to:
• Inspect model outputs
• Compare predictions vs. ground truth
• Explore dataset slices
• Debug failure modes
• Build review and triage systems that streamline model analysis workflows for CV researchers.

Applied Computer Vision Research Support

• Collaborate with research teams to analyze model performance on high-resolution satellite and aerial imagery.
• Help identify systematic failure modes and propose evaluation-driven improvements.
• Support experimentation with state-of-the-art CV models including foundation models, segmentation networks, and geospatial ML architectures.

Cross-Team Impact

• Establish best practices for evaluation and ML experimentation across the organization.
• Provide technical leadership in reproducibility, benchmarking, and model validation.
• Mentor engineers and researchers in designing reliable evaluation systems.

Required Qualifications

• 7+ years of experience in machine learning, computer vision, or applied AI engineering.
• Strong experience in computer vision model development and evaluation.
• Experience working with satellite, aerial, or geospatial imagery (or similarly large-scale visual datasets).
• Proven experience building ML evaluation harnesses or benchmarking systems.
• Experience publishing experiment results to a centralized tracking system (e.g., Weights & Biases, MLflow, or similar).
• Experience building internal developer tools, dashboards, or analysis systems.
• Strong programming skills in Python and familiarity with common CV/ML frameworks (PyTorch, TensorFlow, etc.).
• Experience designing metrics, ground truth schemas, and evaluation protocols for ML systems.

Preferred Qualifications

• Experience in the geospatial or Earth observation industry.
• Familiarity with remote sensing data formats and geospatial libraries (e.g., raster data pipelines, GIS tools).
• Experience working with large-scale imagery datasets (terabytes to petabytes).
• Experience evaluating foundation models or multimodal vision models.
• Familiarity with Weights & Biases (W&B) or similar experiment management platforms.
• Experience designing model debugging or annotation review interfaces.
• Experience working in research-oriented ML teams.