At Yenda, I annotate video and image datasets against detailed taxonomies to support AI model perception and reasoning. I also evaluate AI-generated outputs for conversational accuracy, semantic alignment, and instruction-following.
I flag model limitations and edge cases with structured feedback, and adapt to evolving project guidelines while maintaining accuracy under deadlines.
At Werkit Zambia, I transcribed conversational audio and performed quality assurance on speech datasets used to train NLP models. I used Jira and Slack to coordinate progress and resolve guideline questions.
