ML Output Reviewer & Transcription Corrector | BHL Field Journal Project, Zooniverse
I reviewed and corrected AI-generated text transcriptions of digitized historical field journals. Using machine learning handwriting text recognition (HTR) outputs, I identified and resolved transcription errors. This process improved the accuracy and reliability of the final dataset. • Worked with outputs produced by Amazon Textract. • Mirrored standard professional workflows in AI model evaluation. • Provided corrected training data for future model improvements. • Supported biodiversity and scientific record-keeping projects.