Uzbek OCR Annotation Specialist

I’m building a multi-type OCR dataset for the Uzbek language and need help drawing precise bounding boxes around every text region in 20 000 scanned images, then transcribing each snippet exactly as it appears. All tasks will be performed in our web-based annotation tool; no local installs are required. You’ll be paid per output, at $0.003 USD for every correctly placed box plus $0.001 USD for every verified character you transcribe. The full set must be finished by the end of September, so availability and consistency are important. What I value most is evidence that you’ve handled image bounding-box annotation before, even if you’re new to large-scale OCR work overall. When you reply, please attach or link to past projects that highlight this skill so I can gauge quality and speed. Accepted files will: • follow the project’s box-naming and layer conventions • contain zero empty or overlapping boxes • match transcription guidelines for spelling, punctuation and diacritics Native-level Uzbek fluency is essential; familiarity with AI training datasets is a bonus. If the terms, timeline and pay structure suit you, send your past work and a short note on your current capacity, and we can get started right away.

Project ID: 39737600