Computer Vision Data Scientist

Join to apply for the Staff Computer Vision Data Scientist role at Line Vision

Join to apply for the Staff Computer Vision Data Scientist role at Line Vision

Remote-Hybrid:
Boston, MA Headquarters or Boulder, CO Hub Office

Empower the clean energy transition by leading cutting-edge computer vision innovation in real-world, mission-critical environments for the global electric grid.

We are seeking a Staff Computer Vision Data Scientist to join our mission-driven team and lead the development of advanced computer vision capabilities across our product ecosystem. This role blends deep technical expertise with mentorship and innovation to expand the frontier of what’s possible in real-world computer vision—particularly in complex, uncontrolled outdoor environments. If you are looking to contribute to a high-impact, fast-paced company where your work accelerates the deployment of clean energy technology, join us at Line Vision, Built In Boston and Colorado Best Places to Work!

What will you do?

Core Responsibilities And Key Performance Objectives

• Architect and own a sustainable, end-to-end machine learning training pipeline for CV models.
• Champion and deliver edge-deployable CV models that are robust, efficient, and optimized for low-power environments in remote areas.
• Mentor and coach data scientists and engineers in CV fundamentals, best practices, and emerging techniques.
• Develop, deploy, and maintain multiple CV models addressing a range of environmental and sensing contexts by gathering and curating novel ML training sets.
• Lead strategic initiatives involving multimodal data fusion (e.g., combining satellite, drone, LiDAR, multispectral, low-light, and RGB camera data).

• Fully onboard and assess current ML training pipelines and CV infrastructure.
• Begin developing a roadmap for a robust, sustainable training system.
• Build relationships across Data Science, Software, and Sensor Hardware teams to understand pain points and opportunities.

• Deliver a working MVP of a model training pipeline tailored to Line Vision’s data constraints and goals.
• Guide team members through project-based mentoring and technical knowledge transfer.
• Pilot an early-stage edge-deployable CV model for field testing.

• Stand up and document a scalable, production-grade CV training and deployment system.
• Ship multiple robust CV models in production, including at least one optimized for edge processing.
• Build a track record of successfully translating general product requirements into actionable tasks and clearly defined model performance metrics.
• Develop best-practice guidelines, model accuracy benchmarks, and reusable assets for future CV models.

• Delivering Innovative Solutions:
  Design novel systems and scrappy workflows to generate and curate fit-for-purpose ML training and testing imagery.
• Empowering Others:
  Level up overall team capability through mentorship, paired work, and shared learning.
• Judgment & Decision Making:
  Make technically sound choices that balance experimentation with performance constraints.
• Stakeholder Management:
  Bring clarity and alignment on complex CV topics across technical and non-technical stakeholders.

• Deep experience building new CV systems from the ground up—not just optimizing components of legacy pipelines.
• Expertise in working with uncontrolled environmental imagery (e.g., outdoor, variable lighting, occlusion-heavy, and complex natural backgrounds).
• Strong knowledge of both classical and machine learning CV techniques, including feature matching, segmentation, edge detection, pose estimation, etc.
• Proficiency in Python and tools like OpenCV, scikit-image, Roboflow, and anomalib.

• Creativity and practicality in building training datasets, including the use of public datasets, synthetic imagery, and domain adaptation techniques.
• Experience deploying ML models to edge devices with compute and power constraints.
• Strong physics intuition and ability to reason through sensor phenomena and imaging artifacts.
• Familiarity with AWS Sage Maker, image labeling…