Remote Radiology AI Trainer — Shape Next-Gen Medical Imaging

DataAnnotation; Woodbury (MN); Remote; USD 172,000 - 276,000; Full time

Benefits offered by this job

Flexible, remote work

Global eligibility

Job summary

DataAnnotation is seeking an experienced Radiologist or imaging professional to help train AI models by auditing and interpreting diagnostic images, including chest X-rays, ultrasounds, and nuclear medicine scans. The role is open to radiologists and imaging technologists and emphasizes high-quality, structured reporting that guides AI improvement.

You will document findings, provide reasoned impressions, and flag discrepancies between interpretations to refine future model versions.

Qualifications

• Must be based in one of the eligible countries (US, UK, Australia, Canada, Ireland, New Zealand)
• Residency-trained or board-certified radiologist, or credentialed/licensed imaging technologist
• Hands-on experience performing or interpreting imaging studies
• Clear, well-organized writing and ability to justify reports

Responsibilities

• Review de-identified studies with clinical context
• Write comprehensive, structured reports documenting findings
• Give clear impressions and recommend next steps
• Compare two independent interpretations and log discrepancies
• Identify where model-generated reports miss or misstate information

Skills

Fluency in English

Residency-trained radiologist

Board-certified radiologist

Credentialed imaging technologist

Clear, well-organized writing

Education

Residency-trained radiologist

Board-certified radiologist

Credentialed imaging technologist

Job description

DataAnnotation is seeking an experienced Radiologist or imaging professional to help train AI models by auditing and interpreting diagnostic images, including chest X-rays, ultrasounds, and nuclear medicine scans. The role is open to radiologists and imaging technologists and emphasizes high-quality, structured reporting that guides AI improvement.

You will document findings, provide reasoned impressions, and flag discrepancies between interpretations to refine future model versions.