Driving Scene Caption Analyst (Remote Contractor)

About the role

We are looking for a detail‑oriented Driving Scene Caption Analyst to evaluate and refine AI‑generated descriptions of autonomous‑vehicle sensor feeds. The role combines knowledge of U.S. traffic laws with strong written English to ensure safety‑critical captions are accurate and clear.

Key responsibilities

• Watch 30‑second sensor video clips and interpret the driving scene from a first‑person viewpoint.
• Assess and edit large‑language‑model (LLM) reasoning traces for logical consistency, coherence and adherence to observed data.
• Ground entities, events and maneuvers to scene coordinates and object references.
• Correct driving plans, justifications, maneuvers and speed information.
• Detect and remove hallucinated objects or fabricated events from AI‑generated text.
• Validate that captions match the actual trajectory, maneuver and speed while applying U.S. traffic regulations.
• Maintain a zero‑error standard under time pressure.

Required profile

• Documented driving experience in the United States with solid knowledge of U.S. traffic laws.
• Professional fluency in English with excellent written clarity and precision.
• Experience in roles that demand precise writing and structured analytical thinking.
• Ability to evaluate multi‑step logical reasoning and maintain high attention to detail under deadlines.

Required skills

• Familiarity with LLM‑generated content and typical failure modes such as hallucinations and reasoning inconsistencies.

What we offer

• Contract position lasting 12 weeks.
• Flexible commitment of 4–40 hours per week with at least 4 hours overlapping Pacific Standard Time.
• Opportunity to work with cutting‑edge autonomous‑driving data and AI research teams.