Computer Vision Expert at Mercor

Computer Vision Expert | $80–110/hr | Remote (US Only) | ~20 hrs/week

Join a leading AI lab's GenAI team as a ground-truth expert, designing and evaluating frontier computer vision models. This is a structured part-time W-2 position (~20 hours/week) employed through Cincinnatus LLC, fully remote within the United States.

Key Responsibilities

• Task Design & Development: Create challenging, real-world computer vision problems targeting capability gaps in frontier AI models — spanning object detection, segmentation, recognition, multimodal/vision-language reasoning, and image/video generation.
• Solution Generation: Build out problems in an agentic development environment, preparing all necessary components using Python.
• Model Evaluation & Analysis: Assess model performance on your designed tasks and document findings clearly.
• Headroom Identification: Identify failure cases and classify the nature of model shortcomings.
• Expert Collaboration: Work alongside fellow subject-matter experts to ensure evaluation consistency and accuracy.

Core Qualifications

• Deep, hands-on experience in computer vision through industry, research, or graduate/PhD work.
• Working proficiency in Python (applied, not theoretical).
• Strong command of modern CV methods including deep learning architectures and standard tooling (PyTorch/TensorFlow).
• Ability to commit reliably to approximately 20 hours per week.
• Prior experience in AI training, model evaluation, or data annotation is a plus.
• Excellent written communication and strong self-management skills.

About the Role

This position is a structured, role-based engagement — not freelance or project-based. Employment, payroll, benefits, and compliance are administered by Cincinnatus LLC, an enterprise staffing company placing skilled professionals within leading technology client teams.

Cincinnatus LLC is an Equal Employment Opportunity employer committed to diversity, inclusion, and providing reasonable accommodations for qualified individuals with disabilities.