Hackweek XL

Engineering

Learn from our challenges and triumphs as our talented engineering team offers insights for discussion and sharing.

Hackweek XL

Engineering

LiveRamp started the new year off strong by hosting Hackweek XL. Besides being the most aesthetically pleasing roman numerals we’ve had in years, our 40th Hackweek was our biggest yet. As we’ve recently surpassed 300 employees, one of our priorities was to make sure that all LiveRampers felt included and able to participate in the hacking. To encourage that, we sent five San Francisco engineers to bring the Hackweek spirit to the Arbor (now LiveRamp) engineers in NY. Also, while the projects featured below will be engineering-centric, many other departments including sales, marketing, and product presented their ideas.

 

Go Links

Go Links is an internal url aliasing system that aids in knowledge sharing. For instance, when the user types go/hackweek into the browser, they are redirected to a page containing all the information about Hackweek. In a great example of cross team collaboration, LiveRampers partnered with the Acxiom Workspace Experience team to develop an execution plan. We decided that Acxiom would build the backend service and manage the DNS settings while LiveRamp focused on the internal web app. Members of both our full stack and backend teams were able to get Ruby on Rails experience by spending Hackweek building the UI to view, add, and edit Go Links. They also started development on the Slack integration. We’re looking forward to using Go Links to improve communication and discovery of some of our most frequently used docs and support pages!

JAR from Afar

Running exploratory (we know them as one-off) scripts is usually an iterative process. You write a script, test it, add a couple of lines, run it, realize you got one or nine things wrong, make some changes, etc. This is fine when running scripts locally, but when these scripts run on our Hadoop cluster, we have to repackage and deploy after every change. This process can take anywhere from 2-6 minutes. JAR from Afar brings that time down to 30 seconds or less in a majority of the cases by updating the remote scripts directly. The payoff is illustrated in the sample projects below (the new method is update-jobjar). Faster development makes for happier engineers, so we were all very pleased with the outcome of this project.  

DocuRamp

The DocuRamp team approached this Hackweek by trying to capture common issues contributed by the lack of documentation, such as new-hire onboarding and Product-to-Engineering discussions. Their main objective was to then figure out a roadmap and solution for these documentation issues. One such issue we focused on is the many individual data silos that exist in our system (e.g. Dropbox, Google Drive, Box, Confluence Wiki, Flowdock/Slack). They knew that improving documentation shouldn’t stop at the end of Hackweek, so they are continuing to have regular meetings and recently began discussion around improving documentation specifically for new-hire onboarding.

Rampy2.0

cently, all of Acxiom started using Slack as our primary chat platform. This enables employees across divisions to easily communicate and have a single place to search past chats. Unfortunately, we had to move our beloved Rampy chatbot over to the new platform as well. Rampy helps us with many day to day tasks including pinging an on-call, logging ping pong scores, and checking datadog stats. We took this opportunity to start fresh and rewrite the bot from scratch. In the past we used hubot to connect to our chat platform and host the custom scripts. Switching from hubot to slack-ruby-bot enables us to write our scripts in Ruby rather than JavaScript/CoffeeScript. This makes customizing Rampy more accessible for many LiveRampers who work with Ruby on a daily basis. Rampy is now up and running in our Slack instance and it’s very excited about its new home.