• Converting multiple MapReduce jobs to Spark jobs. 
• Updating Spark jobs to use DataFrames and Datasets instead of RDDs. 
• Optimizing existing Spark jobs. 
• Creating Airflow DAGs to schedule data processing.

• Designing and implementing a feature store for machine learning. Preparing a framework for efficient calculation of thousands different aggregate values (features) from terabytes of data. 
• Dockerizing Spark applications to run them as containers on EMR cluster in more isolated and standardized way. 
• Implementing application for serving machine learning model as REST API on Kubernetes cluster using Flask and Gunicorn. Significantly improved response time of the API by using approximate nearest neighbor search algorithm. 
• Implementing lambda function for transforming new objects created in S3 bucket and storing records in DynamoDB table. 
• Implementing REST API application using Akka HTTP that serves recommendations stored in DynamoDB table. 
• Optimizing existing Spark applications. 
• Working with data scientists on optimizing their solutions and making them production ready

• Designing and implementing a feature store for machine learning. Preparing a framework for efficient calculation of thousands different aggregate values (features) from terabytes of data. 
• Dockerizing Spark applications to run them as containers on EMR cluster in more isolated and standardized way. 
• Implementing application for serving machine learning model as REST API on Kubernetes cluster using Flask and Gunicorn. Significantly improved response time of the API by using approximate nearest neighbor search algorithm. 
• Implementing lambda function for transforming new objects created in S3 bucket and storing records in DynamoDB table. 
• Implementing REST API application using Akka HTTP that serves recommendations stored in DynamoDB table. 
• Optimizing existing Spark applications. 
• Working with data scientists on optimizing their solutions and making them production ready

• Implementing report generators for Core Banking Platform using Spark.
• Implementing random data generators for the purpose of verifying the performance of Spark applications. Analyzing outputs of performance tests and making necessary improvements.
•  Implementing Spark jobs for file compaction and repartitioning to improve performance of report generators and Hive queries.
• Working on migration from Cloudera to MapR distribution for Hadoop.
• Using Flume for reading messages from Kafka, transforming them and persisting into HDFS and HBase.
• Using Sqoop to ingest data from Oracle database into HDFS.
• Implementing ETL jobs for transforming files in various formats into Avro format.
• Automated deployment of applications using Ansible which greatly reduced the number of issues during production deployments.

• Using PySpark and GraphFrames to run graph algorithms. Comparing the performance with Neo4j. 
• Using PySpark to transform data stored in S3 and generate CSV files in order to import them into Neo4j.

• Developing application for collecting risk data from various sources and processing it in real-time.

Developing system that uses data from public sources to conclude "who-knows-who" relationships and help companies to identify valuable relations within their existing customers.

• Designing and implementing algorithm for concluding "knows" relationships between persons using Spark.
• Designing and implementing algorithm for finding ultimate beneficial owner of company using Spark GraphX.
• Creating Neo4j Server plugin for finding shortest paths between nodes in the graph using defined business rules.
• Implementing REST services that perform Cypher queries in order to retrieve data from nodes and relationships.
• Implementing fast data import to Neo4j database by writing directly to the files using batch inserter API.
• Implementing transformations of data stored in Cassandra using Spark into the format that can be easily used to import data into Neo4j database.
• Designing and implementing synchronization between Cassandra and Neo4j using event driven architecture.
• Implementing searching nodes in the graph using Cypher queries and Lucene index.
• Data modelling.
• Configuring and tuning Neo4j database.

• Developing PSIcarlos system for optimal planning and precise balancing of crude transportation
• Designing and implementing new system functions based on defined requirements. 
• Preparing technical documentation. 
• Close cooperation within international team. Discussing customer?s requirements. 
• Technical support for system users. 

• Developing (mainly optimizing) service for automatic text recognition from scanned documents and retrieving key information from documents using regular expressions. 
• Developing application for documents scanning and barcodes recognition. 

• Designing scalable system architecture. 
• Developing multithreaded WCF services. 
• Implementing calling in different modes by sending requests and handling events sent using the AMI protocol from the Asterisk PBX. 
• Implementing automatic calling by using integration of the Asterisk PBX with the PostgreSQL database to create dynamic call queues. 
• Developing the predictive dialer algorithm that calculates the number of calls to be made based on collected statistics, e.g. percentage of received calls and talk time. 
• Preparing the mechanism of sending, mixing, compressing and saving recorded calls to the database. 

• Implementing advanced plugins for detecting problems and real-time monitoring of devices and applications such as: 
  • PostgreSQL and MySQL database 
  • Active Directory service 
  • VMware ESX servers and virtual machines 
  • .NET applications 
  • Windows and Unix workstations 
  • Cisco, MRV and Juniper routers and switches 
  • APC UPS devices 
  • Devices with undetected type 
• Creating module for monitoring changes in software installed on detected devices.