I was the Hunt Post-Doctoral Fellow in 1996–1997; my Hunt fellowship led directly to a successful application for an NIH R01 post-doctoral fellowship with the same mentor, which led, in turn, to my current position as Professor at the University of Illinois. As part of my Hunt post-doc, I rented a very low frequency electromagnetic field meter, and demonstrated that the articulograph presents minimal risk to human subjects; this was published as my first article in JASA. As part of my NIH post-doc, I acquired MRI of static vowel configurations (http://isle.illinois.edu/sst/data/mri/), and demonstrated that a PARAFAC rank-3 factor analysis of the English vowels supports two speaker-independent factors matching the distinctive features. As part of my research in Urbana, I have developed automatic speech recognizers using distinctive feature models grounded in both speech production and speech perception. In 2015, Preethi Jyothi and I published the idea of mismatched crowdsourcing: an American listening to Uyghur is modeled as a machine who transcribes each distinctive feature shared by the two languages, with an error probability related to the similarity with which the two languages use the feature.