Drimva

Meta

Lead, Product Content Engineering

$193K–$269K

Los Angeles, CA +3 · Full-time · Posted Oct 1, 2026

Location
Los Angeles, CA
Menlo Park, CA
New York, NY
San Francisco, CA
Experience
10+ years in video production, content strategy, editorial standards, media evaluation, or relevant fields.
Education
Bachelor's degree in a related field, or equivalent practical experience

About the job

Leads human and automated evaluation of generative image and video models, setting quality standards and translating evaluation results into training priorities and launch recommendations.

What you'll do

  • Design and maintain image and video evaluation rubrics; run human evaluation rounds that provide training and fine-tuning signals.
  • Set launch quality bars, calibrate rater pools, and recommend whether generative media features are ready to launch.
  • Partner with research and modeling teams to turn evaluation findings into model priorities and validate automated measures against human judgments.

What they ask for

  • 10+ years of experience in video production, content strategy, editorial standards, media evaluation, or relevant fields.
  • Experience authoring video and/or image quality evaluation rubrics and evaluating generative media output.
  • Hands-on video production expertise and experience partnering with AI/ML research or modeling teams.
  • Experience running human evaluation programs at scale, including rater calibration and inter-rater reliability.

Nice to have

  • Experience building an evaluation function or quality framework from zero-to-one.
  • Familiarity with generative media tooling, annotation platforms, eval tooling, or data pipelines supporting model training.

Pay

$193,000/year to $269,000/year (base salary)

Tools

LLM-as-a-judge, auto-metrics

AI focus

Evaluates and improves generative image and video models, provides training and fine-tuning signals, and scales human evaluation into automated or LLM-as-a-judge measurement.

Model-written summary · the original posting is authoritative

First seen Oct 15, 2026 · Last verified Nov 5, 2026 (UTC)