Data Scientist
2019 - PRESENTDelen Private Bank (via Exellys)- Analyzed fraudulent transactions and created a classification model to detect fraudulent transactions (financial transaction analysis).
- Created a model to segment customers into different groups based on data (customer behavior analysis). This enabled the company to better target customers with relevant events.
- Set up a data warehouse and provided technical follow-up of the development where data from different operating systems are merged together for reporting purposes.
- Designed the model of the data warehouse where data from the different operating systems are merged together for reporting purposes.
- Created operational, management and executive dashboards to support the company.
- Constructed a model to predict when a customer is about to churn and recommend appropriate action to management to prevent the customer from churning (customer churn model development).
Technologies: IBM DB2, Microsoft SQL Server, Microsoft Power BI, Python, SQLETL/Data Warehouse Developer
2017 - 2018Belfius Insurance (via Exellys)- Automated the generation of financial documents which contains data from different data sources.
- Implemented the ability to quickly detect corrupt input data in financial reports; corrupt data is a big problem when reporting to the national and international authorities due to the loss of precious time. The company now detects corrupt input data 50% faster and saves a lot of money.
- Constructed a generic data quality framework which allows the company to improve the operational systems and fix the bugs which are found during the process of analyzing the data.
- Designed a generic data quality framework including the high-level architecture and data model.
- Created the design of management and operational dashboards to provide the company with qualitative insights and improve the overall business.
Technologies: SAS Visual Analytics, SAS Data Integration (DI) StudioData Engineer (Internship)
2017 - 2017Big Industries- Built a Yum repository server which allows one server to function as the only with an internet-connected node in the network of the data center and provide the rest of the nodes with packages and updates.
- Installed and protected (Kerberize) a Hortonworks cluster.
- Let TensorFlow run on a Kerberized cluster using Apache Slider to distribute the computation of AI models.
Technologies: CentOS, Linux, Apache Spark, TensorFlow, Hadoop