Document and Email Data Extraction Reviewer

About OpenTrain

OpenTrain is the #1 platform for people who build careers in AI training and data labeling. We connect skilled contributors with real projects that teach AI systems how to read, understand, and structure human information.

As the hiring and contracting organization for this role, OpenTrain manages onboarding, work assignment, and payments for contributors worldwide.

Why AI training matters

AI training is the human side of creating reliable models: people annotate, correct, and evaluate examples so models learn to behave correctly. Document extraction is a core task — your work directly improves how AI turns messy documents into accurate structured records.

This is flexible, remote work that can fit alongside other commitments. Many contributors begin with entry or intermediate tasks and gain experience that leads to more specialized projects.

The role

You will review records produced by an AI extraction system for documents such as invoices, contracts, bills of lading, certificates, and other business paperwork. Each job contains one or more extracted records with fields like dates, amounts, names, codes, and more.

A single job may include multiple document types and varying extraction schemas. Always read the field names and descriptions for each job — field meanings and expectations may change between document types.

What you'll do

• Compare the AI’s extracted record against the original source document (PDF, email, or scanned file) and confirm fields that are correct.
• Correct fields that are wrong and fill in any fields the AI missed, entering accurate values into the record.
• Flag or escalate ambiguous, missing, or unreadable information according to the project's guidelines.
• Follow the extraction schema and field descriptions closely — do not assume identical field meaning across document types.
• Perform evaluation-style judgments about the AI’s output quality as part of the job (label types include evaluation rating and data collection).

Requirements

This is an intermediate-level task that requires accuracy, attention to detail, and comfort reading business documents. You must be able to work independently as a contractor and follow project-specific instructions.

• Experience with back-office or middle-office documents (invoices, purchase orders, sales orders, bills of lading, certificates) is ideal.
• Previous data labeling or annotation experience — especially with LLM products or extraction tasks — is a strong plus.
• Familiarity with commodities, logistics, or supply-chain terminology is helpful for this project.
• Reliable internet, ability to view PDFs/scans and enter corrected values, and good written communication for escalations.

Compensation & schedule

Pay is hourly at USD 12 per hour. This role is part-time contractor work with a time expectation of less than 20 hours per week.

Open to contributors worldwide; you choose hours subject to project availability and deadlines.

• Payment type: pay-per-hour (USD 12/hr).
• Employment type: contractor, part-time.
• Time commitment: less than 20 hours/week; schedule depends on available tasks.

How to apply and what to expect

Create or update your OpenTrain profile and apply to the project. If selected, you will receive project-specific instructions and field descriptions to guide reviews.

Work is assigned through OpenTrain; follow the provided guidelines, submit corrected records, and flag issues for escalation when necessary. Quality checks and feedback help you improve and qualify for more projects.

• You will be contracted and paid by OpenTrain.
• Project assignments include clear field definitions — read them carefully before starting each job.
• Successful contributors who demonstrate accuracy and domain familiarity may gain access to similar, higher-volume projects.