Job vacancy of Java developer in a/b platform (Vacancy column)

I’ve been writing about engineering managers lately, but we need not only hand-drivers, but also cool engineers for complex tasks. Today I wanted to talk about the position of Java developer in the a/b platform in the team, which makes the engine to count all metrics on experiments for the whole company. Cool if this engineer is senior level and an additional plus will be a deep knowledge of Apache Spark.

If we talk about the a/b platform, then this is a really important project for the whole company, which carries engineering complexity. (I'll tell you more about her.). The A/B test is a classic book by Ron Kochavi.Confidential a/b testing"that I mean." told earlier. If we talk about the history of a/b platform, we had it a long time ago, but before there were a lot of them and they piecewise covered different levels of tests - when I came to the company order. 9 Years ago, there was a prototype of the DCO platform, which we have since rocked and used for experiments in the web and mobile applications. A couple of years ago, we decided to bury DCO and make a new Hippo platform. (I did. pointed her potential logo).

Since then, the platform has grown a lot and learned a lot, but it’s time to scale it to the entire company and to do this, you need to do a lot by calculating the metrics that went the next way. All product analysts count experimental metrics in their python laptops (It is easy to make mistakes or get results by trying data from different angles.) The Laboratory of Experimental Statistics wrote its algorithms for calculating metrics for a/b experiments on the same labs, but for a limited set of metrics. The engineering team at the metric store took these python laptops and transferred them to Apache Airflow, where the calculations were conducted on top of Greenplum, where the MPP database was our main data store for analytical computing. The metric store engineering team has moved the calculation of these metrics to a new architecture, where metrics are counted using Apache Spark, and the data is already inside our Data LakeHouse. (DLH)

In fact, our data platform has migrated to the DLH approach that we somehow did. stripper With Kolya Head on the Research Insights podcast. By the way, in the near future there will be a series of podcasts with an analysis of how we did it with the technical director of the data platform. Plus, there will also be a separate podcast episode about the a/b platform, where we’ll talk about what it is, how important and complex it is. But if you wanted to solve complex technical problems now, you can write to me in person and we will discuss options.

P.S. By the way, I already have episodes.

• With the head of this department Andrei Tsybin, where we spoke About the neighboring product Statist (product analysis)
• With the lead of the backend of this product Zhenya Kozlov, where we spoke both about the Statist backend and in general about the engineering approach to development

#Engineering #Software #Career #Vacancy #Databases #Fintech