Label Shoppable Video Transcripts for Sales Features

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We connect people with projects that teach AI how to understand language, images, audio, and more — with flexible remote work that fits around other commitments.

As part of the OpenTrain community you’ll be contributing to the human side of AI development and gaining practical experience on projects that directly shape model behavior.

About AI training and why it matters

AI training (also called data labeling or annotation) is how humans teach models what to pay attention to. For language tasks this means carefully reading text and marking features so models learn to detect nuance, intent, and factual signals.

This work is often remote, flexible, and accessible: many projects require strong attention to detail and language fluency rather than formal technical credentials, and contributors directly influence how AI systems behave.

The role

We need a contractor to label 10,000 transcripts from shoppable videos on social media. Each transcript must be reviewed and annotated to indicate whether each of roughly 30 predefined video features is present or absent (examples include urgency cues and mentions of discounts).

You will use the project's internal proprietary tooling to record labels. This is a contract, part-time role requiring consistent availability (see schedule below).

• Data type: text (video transcripts from social media).
• Label types required: Classification, Entity/NER + Classification, and Fine-tuning labels.
• Labeling software: internal proprietary tooling provided by the project.

What you'll do day-to-day

Work through assigned transcripts, read each transcript carefully, and mark whether each feature in the provided checklist appears in the text. Expect to make fine-grained judgments about tone, implied urgency, promotional language, and short or subtle mentions.

Follow the project's labeling guidelines and examples closely, ask clarifying questions when needed, and incorporate feedback from quality checks to maintain high consistency across the dataset.

• Review transcripts and label presence/absence for ~30 predefined features.
• Identify and tag named entities or phrase-level features when required by the checklist.
• Apply fine-tuning label formats when indicated for downstream model training.

Requirements

This is an intermediate-level task that requires a very strong command of English to detect nuance and subtle promotional cues. You must be able to read carefully and apply consistent judgments across many examples.

All other specifics below come from the project brief and are required as stated.

• Experience level: Intermediate (attention to nuance and consistency required).
• Strong command of written English — able to detect tone, implication, and subtle mentions.
• Availability: 20+ hours per week.
• Employment types: contractor, part-time (remote, worldwide).
• You will use the project's internal proprietary tooling to complete work.

Compensation, scope, and logistics

Pay structure for this posting is a fixed price of USD 425 for the described work. The project scope is 10,000 transcripts as stated in the brief. There is no hourly rate listed; the fixed-price payment is the only disclosed compensation.

Work is remote and open worldwide. The project will supply labeling guidelines and access to the internal tooling; follow those resources for formatting and quality rules.

• Fixed price: $425 (USD) — as stated in the project data.
• Project scope: label 10,000 shoppable-video transcripts.
• Worldwide applicants accepted; remote work only.

How to apply and what happens next

Apply with a short note about your relevant experience labeling text or working with transcripts, your weekly availability (confirm 20+ hrs/week), and a brief example showing your attention to nuance in English (one or two sentences explaining how you’d decide whether a line expresses urgency).

If selected, you’ll receive the labeling guidelines, access to the internal tooling, and initial test tasks or a small qualification set to verify consistency before full assignment.

• Be explicit about your weekly availability and previous annotation experience.
• Expect a qualification task and feedback loop before receiving large batches.
• Work proceeds in batches through the internal proprietary interface.