Kevin Crooks

Background

About Kevin Crooks

I can't remember a time I haven't been involved in computational linguistics. I have been working in the field since I began my undergrad in 2005, starting as an annotator before becoming a supervisor for 3 biomedical NLP research programs at CU-Boulder. Once graduated, I continued working on those projects while also working part-time for Rosetta Stone on language data annotation. I was later hired by Sensory to work on language models for Large Vocabulary Continual Speech Recognition solutions - our TNL product line. I have worked extensively on statistical language modeling for IoT and telecommunication domains, including medical, automotive, home appliance, and personal digital assistants. I currently manage the Technology Linguistics team. My primary focus has been on ASR - DNN acoustic modeling, lexical systems, and phonology - and NLU - named entity recognition, domain language modeling, and dialogue systems. I have also worked with anaphora and coreference resolution, and syntactic modeling such as PropBank and TreeBank. Through my university research projects, combined with linguistic study of etymology, I have a strong background in medical language. Fluent in English and Spanish. My main scripting language is currently Tcl, but I am comfortable transitioning back to Python after a few years off. Also familiarity with C/C++.