Document + Email Data Extraction

About OpenTrain

OpenTrain (opentrain.ai) is a central job board for data-labeling and AI-training roles. It aggregates positions from many AI companies and labeling platforms so contributors can find this work in one place.

Creating an OpenTrain account is free and applying to a role typically takes only a few minutes.

About AI training and document annotation

AI systems learn from examples curated and reviewed by people. For document projects, human reviewers verify and correct structured data that models extract from files so downstream systems can rely on accurate records.

This posting focuses on document- and email-based extraction: human reviewers compare model output to source documents and ensure fields like dates, amounts, names, and codes are correct and complete.

• Work is fully remote and often flexible in hours.
• Many projects require only attention to detail; domain experience (e.g., back-office documents) can increase your effectiveness and pay.

The role

OpenTrain AI is hiring part-time contractors to review AI-extracted records from documents and emails. You will inspect each extracted record, compare it against the source file, and take one of four actions: confirm, correct, fill in missing fields, or flag ambiguous items for escalation.

Jobs may include a variety of document types (invoices, contracts, bills of lading, certificates, sales orders, purchase orders, etc.), and the extraction schema will change depending on the document type. Always read field names and descriptions carefully — do not assume identical meanings across document types.

• Employment type: Contractor, Part-time.
• Workload: Less than 20 hours per week.
• Location: Remote — worldwide.

What you'll do day-to-day

You will work in a labeling platform (OTHER) to validate structured data the AI produced from documents and email sources. Each task contains the source file and the AI-extracted record to compare.

• Compare the AI’s extracted fields against the source document and confirm correct fields.
• Edit or replace incorrect field values so records match the source.
• Fill in fields the AI missed when the information exists in the document.
• Flag ambiguous, inconsistent, or missing information for escalation.
• Adapt to changing extraction schemas across different document types and follow field descriptions closely.

Requirements

This is an intermediate-level data-review role. Candidates must be comfortable with back-office and middle-office documents and able to follow detailed field instructions without assuming field meanings are constant across document types.

• Experience level: Intermediate.
• Ideal: experience with invoices, sales orders, purchase orders, bills of lading, contracts, or other back-office documents.
• Big plus: prior data-labeling experience with LLM products or similar AI systems.
• Label types: Evaluation/Rating and Data Collection.
• Data type: Document (PDF, email, scanned files).
• Labeling software: OTHER (proprietary or platform-specific tools).

Compensation and logistics

Pay is hourly: USD $12 per hour (PAY_PER_HOUR). Hours are part-time and flexible, under 20 hours per week. You will be engaged as a contractor and can work remotely from most countries.

• Hourly rate: $12 USD per hour.
• Time requirement: Less than 20 hours/week.
• Employment types: Contractor, Part-time.
• Worldwide: Yes — remote contributors may apply.

Who should apply and how it works

Apply if you have attention to detail, experience with back-office documents, and an interest in helping improve AI by verifying structured records. Prior annotation work with LLMs is highly desirable but not strictly required.

To apply, create a free OpenTrain account and submit your application; the process usually takes only a few minutes. If selected, you'll receive instructions for the labeling tool and project-specific schemas.

• Subject matter for this role: Commodities, Logistics.
• Ideal applicants: detail-oriented reviewers with relevant document experience.
• Application: Free OpenTrain account, quick application process.