• Build, validate and productionalize ML models based on tabular, text and image data;• Wrapping up models in REST API (using tornado, docker);• Develop microservices (using kafka, postgres, rabbitmq, pytest, S3);• Build ETL pipelines (using pyspark, hive, hdfs, parquet);• Conducting research and hypothesis testing (using pandas, numpy, scipy, matplotlib, etc.);Stack used at work: Python3, PySpark, Hadoop, PostgeSQL, Kafka, Docker, AWS S3, git, bash, pandas, numpy, sklearn, catboost, pytorch, jupyter

• Conduct ad-hoc research and analysis of clickstream data;• Develop new and improve existing ML models with further deploy to production;• Build ETL pipelines supporting BI & Analytics team;Stack used at work: Python (pandas, numpy, matplotlib, sklearn, pytorch), Vertica, PostgreSQL, Docker, Git, AWS

The main scope of our team was prediction of customer demand. We were using analytical methods to improve overall forecast accuracy for over a dozen of distribution centers worldwide. Along with working with data my role was also to provide tactical recommendations aimed to secure business KPIs which I then presented and validated with stakeholders.Besides that I was taking care of several legacy visualization and data cleaning scripts written in R: maintaining and adapting them for new business needs.Training newcomers and writing guidelines and documentation was also an important part of my work.I was also honored to represent Valeo as a one of two mentors on Final tour of Data Science Game 2017 and help finalists to crunch our data.Stack used at work: R (Shiny, ggplot), SQL, QlikView, SAP, Git, Custom time series forecasting tool, Python (sklearn, pandas, numpy), Tableau

I was focused on optimization of supply chain between several key suppliers and a dozen of warehouses in Europe. Inventory optimization and shortage prediction approach which I put in place resulted in +5pts increase in customer service rate up to 97%.Besides that I dedicated my time to creating various kinds of data visualizations aimed to help the team to get a better grasp of problems at hand.Sharing best practices and training newcomers were also an exciting part of my work on this position.Stack used at work: SAP, SQL, QlikView, VBA, Google docs, R (Shiny)

My role involved managing inventory levels at local warehouse, predicting inventory shortages and close work with international supply chain team (which I joined later in 2014). Data visualization and reporting automation which I put in place were also of a great help supporting quicker problems detection and more robust data driven decisions.Stack used at work: SAP, SQL, VBA, QlikView

Gathered and analyzed data from different modules of ERP systemCreated data dashboards, graphs and visualisationsEfficiency boosting through processes automationStack used at work: Microsoft Navision (ERP), SQL, VBA