I am a Conversational AI and Computational Linguistics leader with expertise in AI-driven automation, NLP research, and product strategy. With a PhD in Linguistics from USC, I specialize in speech and text data research, ASR & TTS systems, noisy query processing, and data curation. My work has been instrumental in developing and scaling AI-powered chatbot solutions, optimizing virtual agent performance, and enhancing automation workflows.
An NLP and GenAI enthusiast, I am passionate about improving AI-driven language understanding for under-resourced languages and leveraging ML and automation to drive customer engagement. I have hands-on experience with Python, SQL, and AI model development, collaborating with cross-functional teams across engineering, product, and business stakeholders.
After teaching at UC San Diego, I transitioned into Conversational AI and NLP applications in industry, working with Google, Apple, Microsoft Research, and Espressive. Most recently, I led Conversational AI operations at Espressive, managing a 14-person team across AI research, data analytics, and chatbot content, and developing AI-driven automation strategies to enhance CX and business impact.
I received my PhD from the Linguistics Department at the University of Southern California in 2016. My linguistic research explores the underlying structure of the human language -- how the structure is coextensive with the meaning. My dissertation is on quantification and ambiguity, analyzed with formal models and psycholinguistic experimentation.
Beyond my career in AI, I am also a dedicated martial artist, training in Karate, which has instilled in me discipline, resilience, and a structured approach to problem-solving. Additionally, I run a handmade jewelry business, where I apply my creative and entrepreneurial skills to design unique pieces and manage business operations.
Humans have two systems of numerosity: one estimates differences in quantity measurement, the other marks and counts distinct individuals through time and space. I investigate this by formal semantic model that accounts for the ambiguity and have implications for processing quantity expressions. Details in my research.
Cross-linguistic research suggests that there are two types of definiteness across languages. I demonstrate that mental activation of the perceived object directly interacts with definiteness construal in languages. Details in my research.
Languages that morphologically mark numerals for counting do not have a systematic plural-marking mechanism. I explore this via syntax-semantic interface. Details in my research.
Salience of an object differs by means of how it is being referred to. My experimental work gathers insights from contextual uses of pronouns and demonstratives, and shows that cognitive coherence has significant effect on reference resolution. Details in my research.