Remote | Senior Computer Vision Assessment Consultant — Up to $150/hour

24-Mag•New York, NY

1d•Remote

About The Position

We are sharing a specialised part-time consulting opportunity for senior computer vision and machine learning professionals experienced in vision foundation models, object identification, image-based quality scoring, defect detection, model benchmarking, feasibility assessment, and executive-level technical reporting. This role supports current and upcoming remote consulting opportunities focused on computer vision feasibility assessment, image model evaluation, baseline benchmarking, data quality review, performance ceiling analysis, production-readiness assessment, and high-quality project execution. Selected professionals will evaluate whether a computer vision system can reliably identify and grade physical objects from images and translate findings into a clear decision-grade report.

Requirements

5+ years of experience in computer vision, machine learning engineering, applied ML, or related technical work
Hands-on experience fine-tuning modern vision foundation models
Experience classifying or grading physical objects from images, including identification, condition scoring, quality scoring, defect detection, or similar use cases
Strong understanding of evaluation design, representative sampling, train/eval separation, accuracy benchmarking, calibration, and validation methodology
Ability to assess feasibility and production readiness of a computer vision system
Strong written communication skills and ability to explain technical findings clearly to non-technical stakeholders
Ability to work independently in a remote, project-based environment
Academic backgrounds in computer science, machine learning, artificial intelligence, data science, electrical engineering, robotics, applied mathematics, statistics, or related fields may be highly relevant
Professional experience in computer vision, ML engineering, applied research, model evaluation, image analysis, or technical assessment may be especially valuable
Equivalent hands-on computer vision and ML experience may be considered depending on project needs

Nice To Haves

Experience with authentication, counterfeit detection, anomaly detection, defect detection, or quality inspection
Exposure to private equity diligence, technical due diligence, feasibility assessments, or other time-boxed advisory work
Familiarity with imaging hardware and capture pipelines, including cameras, lighting, controlled image capture, or dataset collection
Experience with edge deployment, on-prem deployment, production ML systems, or applied computer vision pipelines
Ability to produce clear technical recommendations under a focused project timeline

Responsibilities

Assess the feasibility of a computer vision system designed to identify, classify, or grade physical objects from images
Evaluate whether model performance is strong enough for practical use based on available data, task complexity, and expected accuracy standards
Assess data quality, image quality, label quality, class balance, edge cases, and realistic performance ceilings
Identify technical gaps that may affect reliability, scalability, or production readiness
Benchmark baseline model performance on a representative image sample
Measure accuracy against a held-out evaluation set using appropriate metrics and validation practices
Apply strong evaluation discipline, including representative sampling, train/eval separation, honest benchmarking, and calibration
Review model performance across tasks such as object identification, condition scoring, quality grading, defect detection, anomaly detection, or similar image-based classification tasks
Translate technical findings into a clear, decision-grade report for a non-technical executive audience
Explain feasibility, expected limitations, data constraints, model performance, and recommended next steps
Document methodology, assumptions, evaluation results, and technical conclusions clearly
Provide practical guidance on whether the system should proceed, be refined, or require additional data and testing