Consultant / Digital Nomad

As Head of NLP I lead the team of computational linguists and software engineers who developed the NLU, NLG, Dialog Management, and various other NLP components of our watsonx Orders software.

McD Tech Labs was an AI center of excellence at McDonald's that resulted from their acquisition of Apprente. We created and tested a voice-based conversational agent for the drive-thru.I hired and lead a team of computational linguists who developed various NLP subsystems, including NLU, NLG, Dialog and UX, pronunciation lexica, G2P, text normalization, and others. I wore several hats: tech lead, engineering manager, and individual contributor, as well as advising on product decisions.

I was part of the original team at Apprente, a startup that built linguistically robust, naturalistic, low-latency voice-based conversational agents. I wore many hats, but one of the larger ones was developing an innovative approach to NLU, where generative grammar-based simulation is used to quickly produce labeled training data in domains where data sets are not available or cannot be kept up to date. I co-authored several patents relating to this approach, and we used it to quickly develop bespoke voice agents for brick and mortar businesses.In October 2019, Apprente was acquired by McDonald's Corporation, our biggest customer at the time, and became McD Tech Labs.

I periodically taught a Computational Linguistics course as visiting faculty. The curriculum was a 50-50 mix of NLP techniques and formal language theory.

- Instructor for graduate and undergraduate courses on Phonetics, Computational Linguistics, and College Teaching of Linguistics (pedagogy)- Teaching Assistant for numerous courses- Grant-funded research projects on phonetics, phonological typology, semantics, and historical linguistics

I was the PM for the US English Speech Data Evaluation team, an in-house team of linguists who created golden sets by performing various linguistic annotation tasks including audio transcription, intent tagging, pronunciation lexicon development for ASR and TTS, and ASR error analysis. I worked closely with localization teams doing similar work for other languages around the globe. My team also assisted with QA for crowd-sourced data annotation pipelines.