Welcome! My name is Lyndon, and I'm a data scientist focusing on NLP and computational linguistics. My interests are in computational modeling and acoustic phonetics; in general, I study the interface of corpus data and acoustic phonetics with artificial intelligence.

Outside of school, I compete for the province of Alberta in skeleton sliding.

My thesis work looked at whether there was any acoustic correlates which cause speech to be perceived as more honest. I discovered that ending your sentences with an upward inflection makes you sound less honest.

My most recent publication showed that artificial neural networks can be used to document vowel space for very noisy speech data (recorded in Toronto restaurants). In this work, we studied the endangered Romance language Faetar.

I am available to consult on a variety of linguistics-related topics, particularly involving computational linguistics or natural language processing.

My educational background in the liberal arts, combined with industry software development and machine learning experience, allows me to approach tasks uniquely and effectively. I aim to provide quantifiably valuable machine learning- and data-based solutions to the very human problems found in modern consumer-facing applications.

My methods typically involve statistical analysis of language using either R or Python. Machine learning-wise, I have industry-level compentence working with Python machine learning libraries including Tensorflow, SciKit-learn, and PyTorch. Finally, I provide human-understandable explanations of my solutions, while explaining specifically how they directly bring value to a company.

Specifically, I have industry-capable skills in the following areas. For consulting inquiries, please email me at mail@lyndonrey.ca.

Text-Based Natural Language Processing:
- Sentiment extraction
- Topic recognition
- Linguistic analysis of social media data
- Computational analysis of large corpora

Acoustic Analysis:
- Machine learning approaches to spoken speech
- Accent and speaker recognition
- Acoustic editing and modification of speech