Computer Vision Engineer
Monarch · Emeryville, California ·
- Work mode
- Onsite
- Employment
- Full time
- Category
- ML ai
We offer opportunities to do your life’s work while helping solve one of the most important technical and moral challenges of our time.
Full-time, in-office in Emeryville, California.
Our mosquito work
Turn raw assay video into precise, reviewable measurements of what mosquitoes do over time. The work begins with detection and tracking, but the scientific outcome is a trustworthy behavioral record that can train and evaluate models.
Key Responsibilities
• Develop and validate methods for detecting and tracking multiple mosquitoes in top-mounted behavioral-assay video
• Derive cumulative landing-zone occupancy, trajectories, spatial distribution, entry and exit rates, dwell time, and other interpretable behavioral features
• Build representative labeled datasets and error analyses across labs, cameras, lighting conditions, arenas, mosquito densities, and occlusion patterns
• Quantify confidence and route uncertain or anomalous results to efficient human review rather than silently producing a score
• Design visual overlays and quality-control tools that let scientists inspect how each measurement was produced
• Work with entomologists and lab teams to improve camera placement, assay geometry, capture standards, and the behavior labels that matter scientifically
Qualifications
• Strong experience with object detection, multi-object tracking, segmentation, pose or trajectory analysis, or related computer-vision methods
• Strong Python skills and experience with PyTorch, OpenCV, or equivalent tools
• Experience building evaluation sets and choosing metrics that reflect the downstream use of a vision system
• Ability to build efficient video-processing pipelines and debug failures at the frame and sequence level
• Clear communication with domain scientists and software engineers
Desired Attributes
• Experience with small-object tracking, animal behavior, microscopy, or other visually difficult scientific video
• Experience with domain adaptation, weak supervision, active learning, or human-in-the-loop annotation
• Familiarity with camera calibration, experimental instrumentation, or cross-site capture standardization
• Interest in making scientific measurements interpretable and auditable
Our crop-protection work
Turn raw assay video into precise, reviewable measurements of what insects do over time. The work begins with detection and tracking, but the scientific outcome is a trustworthy behavioral record that can train and evaluate models.
Key Responsibilities
• Develop and validate methods for detecting and tracking multiple insects in top-mounted behavioral-assay video
• Derive cumulative landing-zone occupancy, trajectories, spatial distribution, entry and exit rates, dwell time, and other interpretable behavioral features
• Build representative labeled datasets and error analyses across labs, cameras, lighting conditions, crop surfaces, insect densities, and occlusion patterns
• Quantify confidence and route uncertain or anomalous results to efficient human review rather than silently producing a score
• Design visual overlays and quality-control tools that let scientists inspect how each measurement was produced
• Work with entomologists and lab teams to improve camera placement, assay geometry, capture standards, and the behavior labels that matter scientifically
Qualifications
• Strong experience with object detection, multi-object tracking, segmentation, pose or trajectory analysis, or related computer-vision methods
• Strong Python skills and experience with PyTorch, OpenCV, or equivalent tools
• Experience building evaluation sets and choosing metrics that reflect the downstream use of a vision system
• Ability to build efficient video-processing pipelines and debug failures at the frame and sequence level
• Clear communication with domain scientists and software engineers
Desired Attributes
• Experience with small-object tracking, animal behavior, microscopy, or other visually difficult scientific video
• Experience with domain adaptation, weak supervision, active learning, or human-in-the-loop annotation
• Familiarity with camera calibration, experimental instrumentation, or cross-site capture standardization
• Interest in making scientific measurements interpretable and auditable
To learn more, visit