About the role
Develop and validate photo-based quality measurements with clear limits, uncertainty, and appropriate human review.
Responsibilities
- Define measurable image-assessment tasks and reliable reference labels.
- Build datasets with documented provenance, permissions, and representative conditions.
- Evaluate models across farms, batches, devices, and operating environments.
- Work with scientists and engineers on uncertainty, monitoring, and quality reports.
Experience and skills
- Computer vision and statistical evaluation
- Dataset design, annotation, and reproducible software
- Clear explanation of measurement limitations
Equivalent practical experience is welcome.
First 90 days
- A dataset specification and annotation guide linked to measurable produce-quality criteria.
- A baseline model and evaluation using independently separated development and test data.
Working at Karpolax
We develop produce-preservation sachets and operate KarpoLink, a platform for fresh produce trade, quality analysis, and batch traceability.
Our operations are based in Uganda. Work location, contract type, pay, and travel requirements will be discussed before an offer.