CLAPPERTONE WEKESA

At Align Turni, I produced Swahili and English audio annotations and transcripts for AI/ML training. I also evaluated AI-generated responses against quality rubrics, flagging hallucinations, mispronunciations, and instruction-following failures.

I compared paired model responses and wrote rationales citing specific rubric clauses. I assigned severity labels to unsafe content and policy violations, then documented recurring failure modes and edge cases to support model retraining.

As a Research Assistant Engineer on the CaP4FF Project, I supported system calibration, optimization, and field validation against technical specifications. I also ran baseline surveys and data collection, and applied quality checks to identify system defects.

In my remote data and quality work with Avid and Expensify, I reviewed and validated expense reports, receipts, and financial data for compliance with company policy. I work best remotely with clear written guidance and shared rubrics, and I raise ambiguous cases early.