Hi, I'm Brandon.

I'm a software developer working for
Hewlett Packard Enterprise.

About

Howdy! I'm Brandon Harrison and I'm years old. I'm a software developer with an emphasis on backend development. My alma mater is The University of Texas at Austin, where I learned how to ride horses and yee-haw real good. But since then I've started wearing Patagonia jackets and complaining about Californian transplants in Denver, Colorado. Right now, I'm working on infrastructure automation at Hewlett Packard Enterprise.

Industry Experience

Education

Projects

The Dad-A-Base https://dad-a-base.online/
Scaling, robust website automatically tested and deployed using Heroku to tell dad jokes
  • Utilizes TravisCI to automatically test and deploy a Gatsby based website. Utilizes a Flask and Postgres based backend to manage traffic with PyTest and Jest to test.
CFB Game of the Week https://cfbgameoftheweek.com/
Automatically ranks college football games by how entertaining they are to watch
  • Invented a ”scheduled-scaling” serverless web app using AWS Serverless Application Model (SAM), NoSQL databases, automated testing and deployment with TravisCI, Jest, and PyTest, and a frontend in React.
  • Created a model using Scikit-learn, SciPy, Pandas, and NumPy in order to automatically rank college football games by an “entertainment score.”
BingoBot for GroupMe https://github.com/BrandonHarrisonCode/BingoBot
A GroupMe bot that generates bingo cards on demand
  • Developed a GroupMe bot that listens for user requests and dynamically generates, stores, and delivers a unique, pseudorandom PDF bingo card using a custom word bank using the GroupMe API, Python3, Flask, Gunicorn, and Heroku.
Shakespeare Machine Reading and Character Analysis GitHub, White Paper
Analyzes Shakespeare’s texts to contextually determine the relationship between characters
  • Led a team of developers in a machine reading project utilizing spaCy, CoreNLP, Postgres, Neo4J, and OpenIE to solve named entity recognition and coreference resolution issues in Shakespearean English to determine relationships between Shakespeare’s characters.
Yelp Sentiment Analysis GitHub, White Paper

Evaluates positive or negative sentiment from Yelp reviews to determine the most and least liked dishes in a restaurant
  • Trained a named entity recognition (NER) and sentiment analysis model to identify dishes and analyze Yelp reviews. Dishes were associated with review sentiments to determine if reviewers had a favorable or unfavorable opinion of their experience after having a dish.
  • Built a frontend in React, with a backend server running Flask, SpaCy, and TextBlob on AWS, supported by a DynamoDB database. Utilized Amazon Turk to crowdsource our training data.
National Parks Service NLP Question Answering GitHub, White Paper

Answers human-written questions about National Parks using Natural Language Processing
  • Independently developed and deployed a dockerized web service to AWS ECS that utilizes DeepPavlov, Flask, Google Maps API, and the NPS Data API in order to contextually answer natural-language questions about America’s National Parks.
Explain a Movie Badly https://github.com/flippedAben/prac-nlp-solr

A website to identify movies based off of one sentence summaries
  • Designed a service to identify movies from one sentence plot summaries from the ExplainAMovieBadly subreddit using a React frontend and an Apache Solr search platform.
  • Constructed a Reddit bot which scraped plot summaries and associated movies, built a search space in Apache Solr, and deployed the webapp on AWS.