I am a Director of AI Engineering at Cotiviti Healthcare Analytics, in San Diego.
Previously I've been at Fidelity Investments, Teradata, and Think Big Analytics. From 2014–2017 I was a research scientist at the German Research Center for Artificial Intelligence (DFKI), also teaching and advising at the University of Saarland.
In 2014 I completed a Ph.D. from the Ohio State University, focusing on scalable language modeling informed by neuroscience.

I enjoy working on language modeling, machine learning, data science, natural language processing (NLP), and related fields. Specifically, this includes GenAI, tree ensemble methods, language modeling, machine translation, and efficient large-scale software for machine learning.

Publications

Software & Data

  My Github Software

• ClusterCat - Fast Word Clustering Software. See also accompanying visualizations
• Tok-tok - Multilingual Tokenizer
• Buckeye Pronunciation Dictionary - like CMUDict, but is based on real data instead of intuitions. It also contains information about frequency and mean length of utterance
• Habeas Corpus - Command-line Corpus Tools
• More

  Corpora

  Persian NLP

• Perstem - a stemmer and light morphological analyzer for Persian
• Persian - English dictionary - 105,000 entries, UTF-8, derived from Wikipedia, under the CC-BY-SA 3.0 (mostly proper nouns)
• Persian Light Verbs - a table of the top light verbs in Persian, based on their frequency in corpora
  

Teaching & Tutorials

• Universität des Saarlandes
• Language Technologies - a graduate core course
• The Ohio State University
• Advising
• Tutorials
• Git FAQ
• PDF Tricks
• Others

     Stuff

• How do I do X in LaTeX?
• acl.bst - An ACL BibTeX style that supports hyperlinks
• lsalikealt.bst - a LaTeX bibliographic style that tries to conform to the Language Unified Stylesheet guidelines. What's great about this one is that if you supply a url entry in your BibTeX file, it will do the right thing™
• osuthesis - a LaTeX .sty file and template .tex file for Ohio State dissertations and theses. It's gotten past Checkpoint Charlie for several dissertations and theses now :-) Look for the trademark tilde-slash-slash on the front page.

Contact Me

Email:   (first name) @𝕕𝕖𝕙𝕕𝕒𝕣𝕚.org     (first name at last name)

It's a crazy world out there. Be curious.

                          –Stephen Hawking