LLM QA Contractor - Image & Text Review

LLM QA Contractor — Image & Text ReviewLocation: Remote; ability to work a consistent schedule aligned to US time zones.Contract type: Hourly contractorDuration: 2 months, with potential extension based on business needs.

About The Team

We develop and validate new AI/LLM use cases that improve the experience for customers,Dashers, and merchants. A key part of our work is ensuring model outputs are accurate, safe, anduseful when grounded in real-world inputs like photos and delivery drop-off instructions.

About The Role

As an LLM QA Contractor, you will review photos and delivery text (e.g., drop-off instructions) andcompare them against LLM outputs to assess factual accuracy, completeness, and adherence toguidelines. You'll apply detailed labeling criteria, flag errors, and provide structured feedback thathelps us iterate quickly. Success looks like consistently high accuracy against a gold standardwhile meeting throughput SLAs.You will report into Kemi Akenzua on our Logistics team in the DoorDash Commerce Platform.You're excited about this opportunity because you will...

• Evaluate LLM outputs against source images and text, verifying correctness (e.g.,
  

landmarks in a photo, apartment entry details, access notes).

• Apply clear taxonomies to tag error types (missing info, hallucination, misread text, policy
  

issues) and write concise rationales.
Maintain a high bar for quality and consistency by following written SOPs and participatingin calibration sessions.

• Track work in our annotation tools/spreadsheets and surface ambiguous cases or
  

guideline gaps to the team.

• Meet or exceed accuracy and productivity targets while handling repetitive review tasks
  

with sustained focus.
Work expectations

• Schedule: :20-40 hours/week; some holiday or weekend coverage may be requested
• Environment: Quiet, secure workspace; reliable internet; ability to use approved devices
  

per security guidelines.

• Compensation & staffing: Hourly, via our staffing partner; classification and benefits
  

determined by the partner.

Requirements

We're excited about you because you have...

• Attention to detail and structured thinking — you can follow detailed guidelines and
  

keep labels consistent case after case.

• Experience with content QA, data labeling, trust & safety, or content moderation
  

(contract, freelance, or in-house), or equivalent skills gained through similar work.

• Strong reading comprehension and visual analysis skills — comfortable interpreting
  

photos (e.g., entrances, signage) and nuanced delivery notes.

• Tool proficiency — confident using web apps, spreadsheets, and annotation/review
  

platforms; quick to learn new internal tools.

• Reliability and confidentiality — able to work a predictable schedule, handle sensitive
  

user-generated content, and follow privacy/security requirements.

• Communication — can write clear error rationales and escalate unclear cases