Aleksei Zinovev

For the last 5 years, I have been developing machine learning and deep learning frameworks (such as KotlinDL and Apache Ignite ML) in Scala, Java and Kotlin, as well as integrating machine learning models with frameworks such as Apache Spark, Apache Hive, Apache Hadoop and Apache Ignite.

As part of the SIG JVM working group, for several years I helped build the Java API for TensorFlow version 2.x.

Based on the Java API for TensorFlow and the Java API for ONNX Runtime, the KotlinDL library is being developed (in fact, it is an analogue of Keras or PyTorch Lightning written in Kotlin).

Also, as part of the Kotlin for Data Science Team at the JetBrains, I took part in the development of a library for vector calculations (Mutlik), a library for working with dataframes, wrote articles (https://zaleslaw.medium.com/), spoke at webinars (https://www.youtube.com/watch?v=PQ8sL-wd0Io) and online conferences, organized work with contributors (more than 20 contributors only in Kotlin DL), coordinated the work of other team members on projects, worked closely with the marketing department and helped organize internships in the company.

Since 2012, when I first launched the Hadoop cluster, I have been interested in distributed computing on Big Data, including training ML models on the large datasets. When Apache Spark appeared on the scene in 2014, I was happy to start using it, because it greatly simplified development, plus tricky optimizations, both at the query planner level and at the data storage level, made it possible to speed up execution dozens of times. I worked in the presale, developed prototypes for customers, created a Big Data Mentoring Program at EPAM Systems, which was attended by more than 300 people in Ukraine, Belarus and Russia, conducted trainings as an independent Apache Spark trainer for the following companies: SberTech, Beeline, Megafon, CROC.

Gradually, the source code of such frameworks as Apache Spark, Apache Kafka, Apache Ignite has become so familiar, like my own, and I decided to start working at GridGain, developing a distributed ML module and integration bridge between Apache Spark and Apache Ignite. Commit after commit, gaining people's trust, I first became an Apache Committer, and then PMC (https://home.apache.org/phonebook.html?uid=zaleslaw). In 2019, the last year before the pandemic, I presented the results of my work at the Spark Summit (https://databricks.com/it/session/distributed-ml-dl-with-ignite-ml-module-using-apache-spark-as-database) in San Francisco and ApacheCon in Berlin.