Sr. Computer Vision Engineer

Brief summary of role:

We're looking for an experienced Sr.

Computer Vision Engineer to own core algorithms for mapping, localization, image analysis, and 3D reconstruction on real-world construction site data.

This role centers on advanced computer vision but integrates modern machine learning methodologies to optimize performance on real-world construction site data.

You will leverage classical CV, deep learning (transformers, multi-modal models), and probabilistic modeling to provide meaningful insights to our customers.

What you’ll be doing:

• Design and architect complex, multi-component computer vision and machine learning systems.
• Develop and optimize image localization, SLAM, SfM, and 3D reconstruction techniques.
• Implement deep learning models for visual perception, leveraging modern techniques (LLMs, VLLMs, transformers, and generative AI).
• Prototype and rigorously test algorithms on real-world construction site data.
• Collaborate with platform engineers to transition research prototypes into production-ready implementations.
• Oversee dataset creation, curation, and annotation to ensure high-quality training and validation data.
• Monitor system performance in production, ensuring scalability and robustness.

What we are looking for:

• M.S./Ph.D. in related field or equivalent industry experience.
• 5+ years of industry experience developing computer vision and machine learning models for image processing.
• Strong expertise in camera models, calibration, and 3D pose representations/transforms.
• Experience with SLAM, SfM, 3D reconstruction, photogrammetry, or VIO.
• Proficiency in deep learning frameworks (e.g., PyTorch).
• Experience working with multi-modal models combining images and language.
• Deep understanding of fundamental CV: feature detection/matching, probabilistic modeling, computational geometry, and numerical optimization.
• Experience with 3D visualizations and debug interfaces (OpenGL/PyQt).

us:

• Be part of a high-growth company shaping the future of ConTech