Computer Vision Engineer

01About the role

DocVerify's core forensic engine relies on advanced computer vision to detect manipulated documents — forged PDFs, edited receipts, tampered bank statements, and GenAI-generated files. As a Computer Vision Engineer, you'll own and evolve the image analysis pipeline that powers all of this. You'll work directly with research and infrastructure to ship models that run at production scale with sub-200ms latency.

02What you'll do

• >Design, train, and deploy image forensic models for compression artifact detection, splice localization, and copy-move forgery identification.
• >Build and maintain the ELA (Error Level Analysis), metadata inspection, and font consistency analysis modules.
• >Develop adversarial evaluation datasets and benchmarks to continuously stress-test detection accuracy.
• >Optimize model inference for low-latency serving (ONNX, TensorRT) across CPU and GPU targets.
• >Collaborate with the ML Infrastructure team on model versioning, A/B testing, and monitoring.
• >Stay current with academic research in image forensics, document analysis, and GenAI detection.

03What we're looking for

• >3+ years of professional experience in computer vision or image processing.
• >Strong proficiency in Python, PyTorch or TensorFlow, and OpenCV.
• >Experience training and deploying CNNs or vision transformers for classification or segmentation tasks.
• >Understanding of image compression (JPEG, PNG) and how editing artifacts manifest at the pixel level.
• >Comfortable working with large image datasets and building data pipelines.
• >Strong fundamentals in linear algebra, probability, and signal processing.

04Nice to have

• +Published research in image forensics, document analysis, or adversarial ML.
• +Experience with ONNX Runtime, TensorRT, or Triton Inference Server.
• +Background in document OCR or layout analysis.
• +Familiarity with GAN-generated image detection or deepfake forensics.
• +Experience building real-time inference systems at scale.