A day in the life of a xukeeper: The RTS team


 

What do you get when you mix a group of passionate engineers with new, cutting-edge technologies and the opportunity to learn from each other? A stellar team! For our next “Day in the Life” Team series, we’re putting the spotlight on dataxu’s RTS (Real-Time Systems) team and how they partner globally to deliver the best to our customers.

Our RTS team is working on the real-time system behind the dataxu product suite. The system processes millions of transactions per second while putting each of those transactions through a series of algorithms, data mining, and data science to match our customers’ advertising campaigns with the best possible inventory. Our RTS system is a highly scalable, distributed system built with a state-of-the-art tech stack. The team enjoys the challenges that come with building a high scale system that runs 24/7 without interruptions or downtime.

RELATED: Spotlight on global collaboration: dataxu’s new hire experience

In this team edition of a “Day in the Life of a xukeeper” meet the RTS team:

Michael Cavedon, Principal Software Engineer
Benjamin Zak, Software Engineer
Oleg Lopatin, Principal Software Engineer
Dave Adams, Senior Software Engineer
Raj Hassani, Software Engineer
Anshul Mittal, Network Engineer
Chandan Javaregowda, Software Engineer
Sandeep Sikka, Systems Architect and Data Scientist
Chandrakanth Bapanpally, Software Engineer
Pratyush Talreja, Software Engineer
Geethanjali Thiruvenkadaswamy, Software Engineer

Keep reading to learn more about the behind the scenes of the RTS team from Dave Adams, Raj Hassani, and Benjamin Zak.

Why did you join dataxu? What did you find exciting? How did dataxu stand out from other organizations in your area?

Adams: I had a few friends and former coworkers at dataxu. I was excited by the opportunity to work with large datasets and massive real-time systems.

Zak: I joined dataxu because of the fascinating technologies, the plethora of growth opportunities, and the incredible people. dataxu utilizes many cutting-edge technologies and offers the opportunity to contribute on a meaningful level. dataxu’s Real-Time Decision Engine is a complex, high speed, and large-scale system. It is a vital component of dataxu’s technology. As such, the RTS team handles challenging and relevant business impacting problems. The team encourages critical thinking and truly gives me the opportunity to address difficult problems with creative approaches.

dataxu is also a great place for personal learning and growth. The RTS team, on an individual level, is highly accomplished and creates an incredible atmosphere for rising engineers. The senior members of the team are always available to advise, direct, and teach. Outside of the RTS team, dataxu is a great place for cross-team learning. As an example, I attended the Optimization team’s “Lunch and Learn” sessions and their weekly presentations. Within a few short months, I developed a strong foundation in Data Science and Machine Learning.

Hassani: I joined dataxu as I wanted to work in a midsize company and dataxu is one of the cool ad-tech companies with a highly modern tech stack. The people here are highly innovative and I saw dataxu as a strong platform where I could both learn and grow my career.

What’s the coolest problem you’ve solved since you started at dataxu? OR what’s your favorite project that you’ve worked on since you joined?

Adams: I developed tools to identify and blacklist fraudulent clicks. While our learning system can optimize for more ads clicked, it’s critical that those clicks come from real consumers, not bots.

Zak: One of my favorite projects I’ve worked on is the design and implementation of low latency pacing metrics. The pacing metrics capture behavioral changes in spend on a per flight basis. They are a powerful tool for debugging any flights that are not spending as excepted and can also be used to evaluate system experiments. I loved working on this project because it involved so many different areas. Throughout the course of the project, I touched high-level design, programming in Java and in Scala, various AWS technologies, and mastering several unique components of the RTS.

Hassani: I worked on a project to add viewability targeting to our platform. This addition helps clients to only show ads on top of viewable slots which helps them drive success in their campaigns.

What technologies are you working with? What’s a new language or skill that you’ve learned since you joined the team?

Adams: Since joining dataxu over 5 years ago, I have picked up Ruby on Rails, Hive, Spark, and raw Javascript. Since joining RTS last year, I have been working with Node.js and Java.

Zak: The Real-Time Bidding System touches many different technologies. Recently, I have worked with various AWS products, such as Kinesis, Firehose, and Lambda functions. I’ve also recently started working with Aerospike, an extremely high-speed big data database. The most valuable skill I am developing on the RTS team is problem-solving. As a large-scale and low latency infrastructure, RTS presents a unique set of challenges. Addressing issues at this scale requires taking a step back to consider some new approaches. How does my solution scale with size? How does my solution affect the speed of the system? Luckily, the senior team leaders do a great job of setting the example for approaching these types of problems.

Hassani: Currently, I am working with Java, Python, SQL, AWS, Postgres, and Qubole.

What is a creative way you partner together as a global team in order to ensure customer success?

Zak: With a strong presence in Ukraine and in India, RTS truly is a global team. We have 24/7 support to ensure customer success.

Have you been involved in any Innovation Day projects? What did you create?

Adams: I have built a dashboard for analyzing errors that happen with video ads, easy-to-run reports for campaign troubleshooting, a monitor for finding and fixing data inconsistency across our subsystems, tools for troubleshooting and fixing creatives, an algorithm to optimize campaigns for profit margin, a platform for quick-to-build monitors and alerts, a tool for previewing creatives across all types of browsers and devices, and an integration between our platform and Salesforce.com to alert dataxu account teams of large changes made by self-serve customers.

Hassani: Last Innovation Day, I was involved in Ask dataxu!, an Alexa-enabled app which will help campaign managers to get their spend and other analysis about their campaigns.

Interested in learning more? Check out our careers page to discover what being a xukeeper is all about.