Cloud architect and DevOps engineer

3/2019 – 12/2020

Green field development of a system for text analysis in legal tech domain. A message-driven microservice system is
built with following technologies - Scala with Typelevel stack (pure FP libraries cats-core, cats-effect, http4s, fs2 etc.),
RabbitMQ, Elasticsearch and modern DevOps tech (Kubernetes & AWS, Docker, Helm, Gitlab). The project language was German.

Amazon Web Services (AWS), Continuous Delivery, Kubernetes, Scala

10/2017 – 1/2019

I worked in Personalization division on a new product - Algorithmic Fashion Companion, that algorithmically creates and recommends outfits. The team went a path from first user interviews to going live on several Zalando applications. I worked on the data and back end infrastructure (AWS, Kubernetes, Akka-Http, Airflow, Spark, Python, Scala). The project was machine-learning driven, so close collaboration with research engineers was part of it.

Apache Spark, Git, Python, Scala, Amazon Web Services (AWS), Kubernetes

6/2016 – 9/2017

I worked in Zalando Logistics in a team, which purpose is developing and operating warehouse software for reliable lifecycle management of hundreds of thousands orders per day. Besides supporting the legacy monolithic application written in Java, the main task was to migrate old and develop new green field projects in a new microservice, event-driven architecture (using Scala, PostgreSQL, Play and Kafka/Nakadi).

Postgresql, Apache Maven, Git, Java (allg.), Jenkins, Scala, Amazon Web Services (AWS)

8/2014 – 5/2016

Social Analytics Platform. The main goal of the Social Analytics Platform is to provide relevant information to telecom companies about their customers (including predicted age group, gender etc.) as well as analytical tools for marketing personnel to conduct marketing campaigns.

This was a new product in the company and my task was to act as a lead developer and shape processes and development culture of a team of 6 data engineers in collaboration with 3 data scientists.

The scope of the project included development of data pipelines and implementation of predictive models on Hadoop stack (Cloudera, HDFS, Map Reduce, Apache Pig, Oozie, Sqoop, Flume, Hive), back end and UI applications as well (Java 8, Spring, Hibernate, PostrgreSQL).

The project had high quality standards via consistent application of best practices - modularity, continuous integration,unit/integration/functional/end-to-end testing.

We implemented several pilot integrations for various customers. As a result we got a first commercial purchase of the product from a major telecom company (over 10M customers).

Apache Hadoop, Projektleitung / Teamleitung (IT), Apache Maven, Git, Java (allg.), Representational State Transfer (REST)

1/2014 – 7/2014

Yandex.Fotki is a photo hosting service with millions of users, one of the media services of Yandex, the leading search provider in Russia.

I got useful experience in development and operations of large-scale, highly available systems in cross-datacenter environment. I did back end refactoring to support the major service redesign and developed integrations with external services. At that time I collected first functional programming experience in production (using yandex-bolts, a Java clone of Scala collection library).

Apache Maven, Git, Java (allg.), Representational State Transfer (REST)

5/2010 – 12/2013

Metamodel is a software company, that mainly focuses on customizing content-management solutions for russian corporations and state departments. The Metamodel platform consists of a number of modules and is built on top of Documentum solutions.

I started as a Java developer and participated in a number of major integrations. Also I built a number of internal libraries and frameworks and eventually become a person in charge of maintaining and refactoring of the code base of core platform modules (600k SLOC). Among my contributions was migrating to a more modern tech stack (SVN to Mercurial, Ant to Maven, introduced Spring).

In 2013 I started a new project as a team lead of 3 developers. The project's goal was to predict optimal schedules for ATMs encashment to cut down significantly banks' maintenance costs. In a year we turned a complex mathematical model in a working software system and proved the economical value for the first customer.

Apache Maven, Subversion, Java (allg.), Junit