Matthew Frost

Summary

Currently a Lead Software Engineer at Disney Streaming heading up the visual personalization team for Disney+. Previously at VMware for over seven years working on Cloud Director.

Before moving into tech I spent a decade in hospitality - primarily in food and beverage management with Hyatt and Lake Compounce amusement park in Connecticut.

I made the jump to tech after almost-accidentally discovering VBA for Excel and finding really enjoyed writing code to solve problems. What carried over from hospitality is that the emphasis on guest experience translates pretty directly to caring about user experience in software - and it's a big part of why building products at Disney has been a particularly good fit.

Skills & Technologies

Most Used/Experienced

Java
Kotlin
Ruby
PostgreSQL
Spring
Hibernate
Rails

Moderate Use/Experience

Go
Lua
JavaScript
MySQL
DynamoDB
Kinesis
AWS
Docker
Kubernetes

Some Use/Experience

C++
Scala
Groovy
PHP
Python
Clojure
TypeScript
MongoDB
Microsoft SQL Server
Oracle RDBMS
Redis

Quickly assessing someone's Twitter activity by glancing at their profile can be tough to get an accurate idea of tweet frequency; shouldifollow was intially created to easily see how often an account tweets and was later expanded to provide other data points.

It displays the following metrics:

Tweets per day
Retweets per day
Most used hashtag
Number of hashtags used
Retweeted by others
Twitter join date
Timing of tweets

View on GitHub.

Working on this is a lot of fun as it allows me to combine my enjoyment of World of Warcraft with my passion for coding. DontCast is an addon that warns players when their target would be immune to damage (i.e. "don't bother casting"). It displays the name and icon of their opponent's damage mitigating ability, along with a timer counting down the seconds until the ability expires.

Originally created for my own use but added to Curse in case anyone else found it useful - I thought maybe a dozen people would download it, but it currently has over 250,000 downloads!

View on GitHub.

As an avid World of Warcraft PvPer I would frequently find myself looking up players on the leaderboards to see what talent choices they were making. This was a time-consuming process which limited the sample-size and so I created PvP Leaderboard.

PvP Leaderboard looks up (daily) the players on each of the leaderboards and then provides the results (with optional filtering) allowing someone to quickly and easily see the talents top players are choosing and view the US and EU leaderboards.

As the name implies this site also lists the players on the leaderboards of each brack in the US and EU regions (wich similar filtering capabilities).

On GitHub: web app (Ruby on Rails), updater (Go).

Received a 3.96 (out of 4.0) on my Capstone Project for WGU, a Java application that predicts stock market movement based on Twitter activity. The program analyzes mentions of a company's ticker symbol to determine a bullish, bearish, or neutral opinion based on the content of each post. Each score is then adjusted for age of post and past accuracy of author before determining the overall sentiment for a stock.