Data Annotator - Fruit Labelling (contract)

About Aerobotics

Aerobotics is a Vision AI company helping fruit producers get more from every harvest. Our models are trained on the world's largest fruit dataset, creating a flywheel that continuously improves the product for our customers. We started with fruit sizing and expanded into forecasting and grading, and we're now developing full yield estimation. Our platform delivers accurate, transparent insights across the fruit supply chain.

To date, we've analysed over 400 million trees from the sky, and more than 300 million fruit up close, helping farmers make critical operational and agricultural decisions that maximise yield and efficiency. Our global team of 50+ professionals is headquartered in Cape Town, with commercial offices in the USA, Australia, Portugal, Spain, and South America.

The Opportunity

We’re looking for a detail-oriented Data Annotator to join us on a 6 month contract in De Waterkant, Cape Town. You will contribute to building out our fruit dataset to boost existing machine learning models and to help us build new ones. The role involves manual interpretation of fruit imagery, and reasoning about the fruit in the imagery: its colour, whether it is damaged, the type of damage. Most of what you’ll label is straightforward, but your job is to make the calls when the decision is not straightforward. If you can spot patterns, reason when things are ambiguous and explain your decisions, you’ll do well.

What you will do:

• Label and categorise fruit imagery: fruit colour, fruit quality, edge segmentation
• Correct existing annotations where something's incorrect
• Flag blurry images, images of non-fruit objects or other erroneous images.
• Assess model outputs and determine the conditions under which it performs poorly.
• Hold a high bar on accuracy. Your labels are the foundation that the models are built on.

What we expect from you:

• Reasons through ambiguous cases instead of guessing or defaulting to “close enough.”
• Takes initiative: flags a problem instead of labelling around it.
• Has the patience for high volume detail oriented work, always maintaining accuracy
• Reliably does their work, and is upfront about disruptions or blockers.
• Drive to expand their technical knowledge and leverage AI and tools to their advantage.

What you’ll need:

• A higher education degree or diploma.
• Comfortable in Excel, Google Sheets, and Docs.
• Your own laptop and stable internet.
• Based in Cape Town, able to get to our De Waterkant office for regular training.
• Available immediately or close to it.

Availability required:

• 6-month contract, with potential for extension based on performance and business needs.
• Mondays - Fridays each week with compensation at R500/full days work
• One week of training at our De Waterkant office to start.
• After that, availability to be in-office Tuesdays and Thursdays for collaboration sessions, and remote the rest of the week

Application Process:

Submit your CV and answer all application questions. Candidates who move forward get a short take-home assessment, followed by a video interview with the team.