This month we enjoyed another successful 24-hour Movio Hackathon, with a whopping 14 teams getting in on the action. This bi-annual challenge is set to encourage the Movio Crew to step away from their usual tasks and spend an entire day and night focused solely on creating something new from scratch. From a ‘Doom’ video game spin-off to social media sentiments, the teams pulled out all the stops to discover what they could whip up - and the results were epic.
As with previous Hackathons, a fair few competitors managed to toughen out the graveyard shift powered by multiple crates of energy drink and 162 large slices of pizza. After 24 hours of eyes being glued to computer screens, the competing teams stepped back into reality and presented their projects to the rest of the office - below are a few favourites.
Team name: Doomernates
Team member: Leo Palmer
Project name: Doomernates
Success rating: 5/5
“Since we have a lot of really flashy tools that make it easier for us to deploy and create our applications, I figured I'd streamline the destruction and decommissioning of our applications.
I decided to use a more tried and true system from 1993; Doom 2. It's a great platform to quickly dispatch of your costly resource-hogging applications.
I started with an open source fork of the original Doom 2 called 'psDoom' which allows you to kill your local processes manifested as ‘Doom’ monsters. I changed this from using local processes to using Kubernetes Pods and Services. It polls the Kubernetes API to get a list of the Pods and Services and then issues a "kill" API call when they are killed in-game. In the above demo, I send those "kill" commands to a text file instead of to the actual API, but the poll API calls are real. What you see in the game are actual Kubernetes Pods and Services manifested as ‘Pink Demons’ and ‘Shotgun Guys’ respectively.”
Project name: Bookie
Team members: Mariano Gappa, Gary Zhu
Team name: Bookie
Success rating: 5/5
“Our goal was to create indexing for Kafka queues. We built a service that scrapes a collection of Kafka topics looking for relevant information, stores that information onto a relational database and exposes it as JSON via an HTTP endpoint. Bookie makes Kafka queues semi-random-access. It’s meant for composability: any tool that consumes a queue can use it.”
Project name: Social Buzz
Team members: Karthik Subra, Sam Stradwick, Jerry Peng, Jack Hopner, Ashleigh Davis, Raghu Kasturi
Team name: The Social Touch
Success rating: 5/5
“Our original idea was to create a dashboard that pulled through all of the ‘buzz’ surrounding current cinema releases from social media, and collate the live feeds into a singl e easy-to-use dashboard.
We decided to focus our Hackathon project on gaining these insights specifically from Twitter. Karthik and Ashleigh researched relevant Tweets and hashtags, Sam envisioned what the final product would look like, and Jerry pulled through data from Twitter to illustrate overall sentiments for a handful of pre-selected movies. Jack and Raghu then converted this data into an interactive widget based off Sam’s ideas.”
Project name: Page ME
Team members: Kalman Bekesi
Team name: Page ME
Success rating: 4/5
“A solution looking for a problem; build a mobile app with React Native. Movio currently use a product called Pager Duty and my idea was to get started on an open-source version that we could use internally.
The goal was to experiment with a number of new technologies and get something up and running with React Native, GraphQL and AWS Lamdba. This was achieved, however, getting it hooked up to DynamoDB would have been the cherry on top.”
Project name: Numero Ads
Team members: Nicolas Maquet, Raghu Kasturi, Matthias “The Magician” Langbein
Team name: Fantastic Ads And Where to Find Them
Success rating: 5/5
“Our goal was to create banners/widgets that could be embedded into websites via a simple JavaScript snippet (similar to Google Ads) and display contextual box office data (e.g. for selected films or for top films of the day).
We used a simple service in Go that retrieved box office data for the top ten films of the week from Numero's console, and embedded it into a JavaScript template that was served to a client. The JavaScript was responsible for targeting a user-defined element on the page and injecting the widget with the box office data for a single film or multiple films, based on context.”
With a minimum success rating of four out of five, it's safe to say we had a particularly triumphant Hackathon this time around. Now that the fatigue and dizziness has worn off, we've already begun planning our projects for next time - so stay tuned!
