Junior AI Captioning Accuracy Analyst

Vordex Tech

Remote · Under $60k / year · Contract · Posted 4 months ago

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

About the role

Our automated captioning system generates subtitles for 40,000 video minutes per month across educational, corporate, and media clients. WCAG 2.2 requires 99% accuracy for accessibility compliance, and we're currently averaging 97.2%. That gap is the difference between a product that's legally compliant and one that creates real accessibility barriers for Deaf and hard-of-hearing users. This is a 6-month contract role focused on analyzing where the accuracy gaps live and building the tooling to close them. You'll review error samples across content categories and speaker demographics, build dashboards that track accuracy by content type, and help the engineering team understand which failure modes to prioritize. The engineers fix the model; you're the measurement layer that tells them what to fix. No ML background required — this is an analytical and quality-focused role. You'll need attention to detail, comfort with data, and genuine care about building technology that works for people with disabilities.

Fully remote.

Responsibilities

Analyze captioning error samples across content types and speaker demographics
Build dashboards tracking accuracy metrics by language, accent, and content category
Define error taxonomy distinguishing critical accessibility failures from minor inaccuracies
Identify accuracy patterns that point to systemic model failure modes
Create evaluation reports for engineering team with prioritized fix recommendations
Maintain ground truth annotation dataset for ongoing accuracy benchmarking

Requirements

1+ year of experience in quality assurance, data analysis, or content evaluation
Strong attention to detail and comfort with systematic error categorization
Basic Python and SQL skills for data manipulation
Genuine interest in accessibility and inclusive technology
Clear written communication for reporting findings to engineering teams