We are looking for a Computer Vision Engineer. Ryz Labs is partnering with a client building an AI-powered platform for automated quantity takeoff in heavy civil and highway construction.

As a Senior Computer Vision / ML Engineer, you'll own the “what is it?” problem—identifying materials, surfaces, and construction features from engineering drawings, PDFs, aerial imagery, and satellite imagery.

You'll work closely with another engineering role focused on geometry and measurement, turning your detections into reliable, priceable quantities.

What You'll Do

Build CV/ML pipelines for engineering drawings, PDFs, aerial and satellite imagery
Extract information from both vector and raster PDFs, including degraded/scanned drawings
Develop detection and segmentation models for materials and infrastructure such as asphalt, concrete, pavement, drainage, pipes, signs, guardrails, and lane markings
Combine information from drawings, specifications, and imagery, identifying discrepancies between sources
Build the ML infrastructure around the models: labeling, datasets, evaluation, regression testing, confidence, and provenance
Work directly with domain experts to create ground truth and improve the system
Prioritize reliability and completeness—“not determinable” is better than a low-confidence guess

What We're Looking For

Required:

5+ years building and shipping production Computer Vision / ML systems
Strong Python and PyTorch experience
Production experience with detection and segmentation on real-world imagery
Experience with aerial, satellite, drone, or comparable raster imagery
Strong knowledge of OpenCV and classical CV techniques
Experience with messy, domain-specific data and human-in-the-loop systems
Understanding of PDF internals, vector geometry, and/or document AI, or the ability to learn quickly
Strong written and verbal English

Nice to have:

Experience with geospatial imagery, construction, document AI/OCR, CAD vectorization, or autonomous vehicle perception
3D experience with photogrammetry, point clouds, terrain/surface models, or cut & fill
Experience across both document vision and aerial/geospatial CV