Senior Computer Vision and Deep Learning Engineer

Quick Facts

• Role: Senior Computer Vision and Deep Learning Engineer

• Focus: End-to-end OCR for scene text and document understanding

• Work style: Remote or office; collaboration on shared codebases and remote projects

Description

Own and improve an OCR pipeline end to end, covering text detection, recognition, decoding, and layout grouping. Diagnose failures on real customer images, develop geometry-aware components (features, homographies, RANSAC, perspective warping), and run controlled experiments with rigorous benchmarking. Work with deep learning and document understanding using PyTorch, OpenCV, NumPy, Hydra, and Weights & Biases.

Responsibilities

• Own and improve the OCR pipeline end to end

• Diagnose failures on real customer images and identify root causes

• Develop and optimize text detection, recognition, decoding, and layout components

• Use image geometry: feature detection/description/matching, homographies, robust estimation (RANSAC), and perspective transforms

• Build and curate datasets; run augmentations, fine-tuning, and controlled experiments

• Benchmark changes against tracked baselines and document measurable improvements

• Collaborate with the team on shared codebases and provide clear technical analysis

Requirements

• Strong programming skills in Python with NumPy and OpenCV

• Hands-on experience training and deploying deep learning models in PyTorch

• Expertise in scene text and document understanding (detection, recognition, decoding, confidence calibration, constrained decoding, reading order/layout grouping)

• Practical multi-view geometry experience (feature detection/description/matching, homography estimation, RANSAC and variants, inlier analysis, perspective warping)

• Ability to debug chained coordinate-space transforms and polygon geometry

• Systematic failure analysis skills (error taxonomy, reproducible test cases, failure buckets, controlled ablations)

• Experience with dataset construction and curation, including augmentation, synthetic/degraded data generation, domain adaptation, fine-tuning, and hyperparameter sweeps

• Ability to benchmark rigorously before and after changes against tracked baselines

• Proficiency with Hydra and Weights & Biases

• Experience with remote GPU servers/clusters and containerized environments

• Experience with Git and GitHub

• Habit of testing numerical and geometric code

• Effective collaboration in team and remote settings

• Clear technical writing, including analysis reports