Computer Vision Evaluation Consultant | Remote

What You'll Be Doing

• Design challenging, real-world computer vision problems targeting specific capability gaps in frontier models spanning object detection, segmentation, recognition, multimodal/vision-language reasoning, and image/video generation.
• Build out tasks within an agentic development environment, preparing all necessary components and reference solutions using Python.
• Evaluate model performance against designed tasks, analyzing outputs to surface reasoning failures and capability limitations.
• Identify and classify failure modes where the target model underperforms, informing headroom analysis and future evaluation direction.
• Collaborate with fellow subject-matter experts to maintain consistency, rigor, and accuracy across evaluations.

What We're Looking For

• Strong experience in computer vision including hands-on work in object detection, segmentation, recognition, or multimodal reasoning from industry, research, or a graduate/PhD background.
• Strong experience in Python, applied practically in research, industry, or open-source contexts.
• Strong experience in modern deep learning architectures and standard computer vision tooling, including PyTorch or TensorFlow.
• Strong experience in model evaluation, LLM/systems training workflows, or structured data annotation is a plus.
• Strong experience working independently in a remote, self-directed environment with clear written communication.

Application Process

Submit your application by completing the application form. Each application will be reviewed against the role requirements. Candidates who meet the requirements will receive an email with the next steps. Follow the instructions provided in the email to complete the remainder of the application process.