Pasha Finkelstein on Big Data, Apache Spark and DevRel
Pasha Finkelstein is a developer, serial speaker, author and host of several podcasts. At the conference Java Meeting Point he will give a talk “Spark: let’s touch it”, in which he will introduce the participants to the world of big data.
In this interview, Pasha told what awaits the conference participants, how he started to engage in DevRel activities and what, in his opinion, should be in a good report.
What are you doing?
I am a Developer Advocate at JetBrains, working on Big Data and Data Engineering. I try to tell people about how the world of Big Data works, what is interesting there, what tools are there.
What will your talk at Java Meeting Point be about?
I will explain what big data is and how it differs from small data. Let’s figure out how Apache Spark works, how its API works, and understand that there is nothing fundamentally complicated about it.
Who will benefit from watching the report?
It will be useful for Java or JVM developers of the Middle + level and for those who are interested in learning how Apache Spark works. We will learn how to write simple pipelines using this framework. It will become clear how, for example, to take and write a data processing pipeline on Apache Spark or analyze data in a dataset.
You do a lot of DevRel activities: give talks, lead podcasts. How did you start doing this?
a friend of mine Slava Semushin went to work in the Czech Republic. I am not particularly good at maintaining relationships at a distance, but Slava is very cool, I wanted to continue to communicate with him. He once wrote, “Listen, let’s do a podcast.” We decided to try, and so was born podcast “Pasha + Slava”…
Once every couple of weeks we called each other and talked about technical topics. Then another one joined us Glory – Artemiev, with whom I worked in the Domklik company. Very convenient – you didn’t have to change the name of the podcast. This is how my first publicity arose.
Then the idea of performing on stage came up, and I gave a talk at the Joker. Subsequently, I revised this report 3 times and delivered it 14 times.
Since then I have been hooked and have given dozens of talks. Over time, it became much easier for me to cook them. I can prepare a speech on a topic that I do not understand in a week, but a topic that I understand in 2-4 days.
Can you advise on how to learn how to prepare performances as quickly?
I guess it doesn’t work that way. All people have their own pace, and you need to love yourself and work in a comfortable mode. And talk about what you want. Reports from under the rod will never turn out good, lived.
What should be in a good report?
It should be lively and interesting, it should have emotions, jokes. There should not be too many thoughts in it, but those that are should not be hackneyed. It is not bad if the information from the report is applied in practice. People become more interested when they understand how they can use this information.
An alternative option is if the report is completely theoretical. On such people will come to twist their brains. This is also possible, but difficult. I once gave a talk about the internals of Kotlin, people didn’t get it.
I like strange formats: live coding, demos, pair reports. I think I’m doing pretty well. But it’s okay not to like any of these formats, choose the one that you like best.
What format will your talk at Java Meeting Point have?
This will be a demo in which a Java or JVM developer will be able to see how big data works using an example of not very big data. We’ll look at slides 5-6 and move on to programming.
