Data Annotation Specialist, AI Training Data

You will work inside Corpshore AI's data annotation and labelling pipeline, the division of Corpshore Solutions Corporation ranked fifth among the top fifty AI outsourcing companies worldwide by Outsource Accelerator. Depending on the project active at the time, you will annotate image, video, text, audio or multimodal data for clients building and refining AI models. The specific task changes project to project, drawing a bounding box around objects in a frame, tagging entities in a block of text, ranking two model responses against each other, transcribing a short audio clip, but the underlying discipline is the same: careful, consistent judgement applied the same way on item one thousand as on item one.

This is a genuinely flexible contractor engagement, not a fixed nine-to-five employee position. You choose your own hours around a minimum of twenty a week, with a modest core window of overlap with AEST business hours reserved for team syncs and guideline updates. It suits people who want substantive, well-paid part-time work they can fit around study, caring responsibilities or another job, rather than people looking for a traditional full-time role.

You will follow detailed, project-specific annotation guidelines and a defined quality rubric for every task you pick up, because the guidelines change by project and a habit carried over from the last one is often exactly the mistake that gets flagged. Your output is checked by a QA reviewer before it is ever delivered to the client, so you are one link in a genuine quality chain rather than the last line of defence, but that also means consistent errors on your part show up quickly. You will flag genuine ambiguity in a guideline rather than guessing your way past it, since one wrong guess repeated across a large batch is far more damaging than a question that slows you down for five minutes. You will maintain your inter-annotator agreement score above the threshold set for each programme, and you will act on the quality feedback your reviewer gives you rather than repeating the same correction request. Depending on the active project you may be labelling training data for a client model, evaluating and ranking model-generated outputs, or verifying transcriptions, and you should expect the mix to shift over time as Corpshore AI's project pipeline changes.

Strong written and spoken English, since the majority of active guidelines and client instructions are written in English regardless of the data type. Genuine, sustained attention to detail, the kind that does not degrade three hours into a shift. A reliable computer and internet connection. The right to work in Australia as a contractor. Comfort with repetitive, detail-heavy work over multi-hour sessions, which is a real and underrated skill rather than a formality on a job ad.

A post-secondary qualification in any field. Prior experience in annotation, transcription, editing, research, teaching or quality assurance. A second language, which opens access to a wider range of projects as they come online. Subject-matter familiarity in any specialised domain, since specialist annotation projects are generally paid at a higher tier.

This is a flexible, contractor-model engagement rather than a salaried employee role, and we would rather be upfront about that than dress it up. What is genuinely on offer: hourly pay that rises with your demonstrated quality tier, so accuracy is what gets rewarded rather than raw speed. Real flexibility over when you work, provided you clear the weekly minimum and show up for scheduled syncs. Work that sits at the front edge of commercial AI development rather than routine data entry. A clear, demonstrated path for strong performers to move into a QA reviewer role, and from there into project coordination, as Corpshore AI's Australian annotation volume grows.