Senior Algorithm Engineer (Computer Vision & Spatial Geometry)

About the Project

We are developing an advanced system for automated analysis and digitization of engineering documentation. Our goal is to transform unstructured visual data (rasters, scans, low-quality PDFs, and plan photographs) into precise, parametric 3D models.

The project combines challenges in computer vision, computational geometry, and high-performance data processing, with a strong focus on pixel- and millimeter-level accuracy.

Role Overview

We are looking for a Senior Algorithm Engineer who will design and implement the core algorithmic engine of the system.

This role is focused on vectorization, topology reconstruction, spatial reasoning, and performance-critical computation, where ready-made ML/CV models are often insufficient in either precision or speed.

You will work on problems where algorithmic rigor, geometry, and heuristics are more important than “gluing APIs”.

Vectorization & Reconstruction

• Design and optimize algorithms for edge and vertex reconstruction from noisy visual inputs (scans, images, low-quality PDFs).

• Handle imperfect, degraded, or ambiguous source data with high robustness.
  

Topology Engine

• Build graph-based engines to interpret spatial relationships between detected elements.

• Automatically infer topology and close geometric contours.
  

Spatial Heuristics

• Implement domain-specific rule systems and heuristics to validate logical and geometric correctness of generated 2D and 3D models.

• Enforce consistency constraints at the spatial and topological level.
  

Contextual OCR

• Design pipelines for extracting textual data (dimensions, technical parameters).

• Integrate extracted information directly with geometric and spatial models.
  

Performance Engineering

• Optimize critical system components for computational complexity and throughput.

• Profile and eliminate performance bottlenecks.

• Use vectorization, parallel computation, and GPU acceleration where applicable.