Computer Vision Engineer

DocVerify; United States; Remote; USD 120,000 - 190,000; Full time

Job summary

DocVerify is seeking a remote, full-time image forensics scientist to own and advance an image analysis pipeline that detects manipulation in documents and images, aiming for latency under 200 milliseconds on CPU and GPU. You will train and deploy CNNs or vision transformers, build error level analysis, and coordinate with infrastructure for versioning and monitoring.

The role requires strong Python experience, PyTorch or TensorFlow, OpenCV, and a solid foundation in linear algebra and signal

Qualifications

• Experience in computer vision or image processing projects.
• Proficiency in Python and deep learning frameworks (PyTorch or TensorFlow).
• Knowledge of CNNs or vision transformers for classification or segmentation.
• Familiarity with JPEG/PNG compression and editing traces.
• Experience handling large image datasets and data pipelines.
• Strong grounding in linear algebra, probability, and signal processing.

Responsibilities

• Train and deploy models to detect compression artifacts, locate image splices, and identify copy-move forgeries.
• Build and maintain error level analysis, metadata inspection, and font consistency features.
• Create adversarial evaluation datasets and benchmarks to test detection performance.
• Optimize inference for CPU and GPU serving using tools such as ONNX or TensorRT.
• Coordinate with infrastructure teams on model versioning, A/B testing, and monitoring.
• Track research in image forensics, document analysis, and synthetic image detection.

Skills

Computer vision

Python programming

CNNs/ViTs

Image forensics

Data pipelines

Tools

PyTorch

TensorFlow

OpenCV

ONNX Runtime

TensorRT

Job description

Role overview

Own and advance an image analysis pipeline that identifies signs of manipulation in documents and other images. You’ll develop forensic models and supporting analysis methods, then prepare them for production use with a target latency below 200 milliseconds. The role is full-time and remote in the EU or US.

Responsibilities

• Train and deploy models to detect compression artifacts, locate image splices, and identify copy-move forgeries.
• Build and maintain error level analysis, metadata inspection, and font consistency features.
• Create adversarial evaluation datasets and benchmarks to test detection performance.
• Optimize inference for CPU and GPU serving using tools such as ONNX or TensorRT.
• Coordinate with infrastructure teams on model versioning, A/B testing, and monitoring.
• Track research in image forensics, document analysis, and synthetic image detection.

Requirements

• At least 3 years of professional experience in computer vision or image processing.
• Strong Python skills and experience with PyTorch or TensorFlow and OpenCV.
• Experience training and deploying CNNs or vision transformers for classification or segmentation.
• Understanding of JPEG and PNG compression and pixel-level traces of editing.
• Experience working with large image datasets and data pipelines.
• Strong foundations in linear algebra, probability, and signal processing.

Nice to have

• Research publications in image forensics, document analysis, or adversarial machine learning.
• Experience with ONNX Runtime, TensorRT, or Triton Inference Server.
• Background in OCR, document layout analysis, synthetic image detection, or deepfake forensics.
• Experience building large-scale, low-latency inference systems.