Mathias Fikru

I’ve been doing AI data labeling and transcription through CrowdGen, annotating, classifying, and validating text data using detailed project guidelines while running quality assurance checks to catch and correct errors. I also transcribed audio into well-formatted text, verifying grammar, spelling, punctuation, completeness, and—when required—timestamps and speaker identification.

Alongside that, I practice text annotation and classification using publicly available datasets, organizing my records in Google Sheets and Microsoft Excel and running my own quality checks to keep labels consistent. I’m focused on reliable, guideline-driven work for AI training and evaluation, especially around NLP and large language models.