At Outlier AI, I evaluate AI-generated responses for accuracy, relevance, instruction following, and overall quality. I compare model responses side by side, classify errors, and document evidence to support quality control.
I also perform Speech-to-Speech evaluation, assessing naturalness, pronunciation, clarity, pacing, and audio quality. I apply detailed guidelines and provide structured feedback to support AI model improvement.
Through my marketing work, including at The 1995 Homestay and as an independent business, I researched customer needs and tested messages using performance data and A/B testing. That experience informs how I assess communication, content, and user needs.
