At Outlier, I conducted comparative evaluations of Korean-language AI models through live speech-to-speech conversations, assessing response content and spoken delivery.
I reviewed pronunciation, prosody, and audio artifacts, and assessed conversational dynamics such as interruptions, backchannels, endpointing, and response latency. I documented turn-level evidence, timestamps, and error patterns to support comparative ratings and quality decisions.
Before this work, I developed apparel collections at Cheil Industries (Samsung) and Kolon Corporation, reviewing samples for design accuracy, fit, construction, and production quality.
