Derek Andersen

Linguistic Engineer

View My GitHub Profile

• • • •

About

I hold a Master’s in Computational Linguistics from Stony Brook University, where I focused on the intersection of language technology and natural language understanding (NLU). My background includes hands-on experience enhancing NLU models and improving their accuracy for voice assistant applications. I’ve developed and optimized models using techniques such as context-free grammars (CFG) in the Java Speech Grammar Format (JSGF) and supported model training through error analysis and accuracy evaluation.

I have a strong technical toolkit, including Python for scripting and tool development, the Linux command line, and Git for version control. I’ve also worked with language modeling approaches like n-grams and probabilistic context-free grammars (PCFGs) and utilized content-based filtering to develop recommendation systems.

Alongside my work in NLU, I’ve developed websites and web applications using JavaScript, Node.js, Express, jQuery, HTML/CSS, Mustache, and MongoDB, further expanding my software development skill set. I’m passionate about applying my skills to new challenges in programming, language technology, and web development.

Education

Year Degree Earned Institution
2019 – 2021 M.A. in Computational Linguistics, Summa Cum Laude Stony Brook University
2016 – 2018 B.A. in Linguistics, Summa Cum Laude Stony Brook University
2013 – 2016 Coursework in Computer Science & Japanese Suffolk County College

Personal Projects

Project Description
Pokémon Solus RGB a romhack of the Gen 1 Pokémon games written in GBZ80 assembly
Headline Snap Toolkit a toolkit for creating and interfacing with a database of news headlines created with Snapchat
Studying Saturdays an automated weekly learning event that takes place in a Discord server
Custom blog engine the backend for my blog, utilizing Node.js, Showdown.js, and Mustache.js
buzzrec (Master’s thesis) a recommendation engine for academic papers in linguistics, utilizing content-based filtering
Language modeling a comparison of two different language modeling techniques: n-grams and PCFGs
banjospeedruns.com the community website for the Banjo-Kazooie speedrunning community
Bingo-Toolie a companion web app for Banjo-Tooie speedrun bingo written with jQuery
PokeQuiz a Python package for command-line Pokémon knowledge quizzes
speedrunb0t an IRC chatbot for Twitch.tv with Speedrun.com API integration

Publications

Coursework