Data Architect /Data Strategist/ Data Engineer / Data Scientist/ Data & AI Management

8/2021 – 11/2021

Creation of a customized Resnet architecture for image classification. Generally, a wide variety of CNN, RNN and Word-Embedding architectures were explored. Using the aforementioned image features vectors, a muti-stage pipeline of models such as XGBoost, CNN, LSTM, custom Restnet and Unet were assembled.

Data Science, Data Mining, Maschinelles Lernen, Neuronale Netze, Amazon Web Services (AWS)

10/2020 – 7/2021

- Setup predictive maintenance cloud Infrastructure with Azure IoT Hub
- Automate stream data transfer from IoT devices to IoT Hub
- Implementation of Messaging Queue to store IoT data in Azure data lake
- Creation of Databricks notebooks and spark clusters for IoT sensor data analysis - Use PySpark , Deep Neural Networks and other Data Mining and Machine learning techniques to analyse data

Big Data, Data Mining, Internet of Things (IoT)

3/2020 – 2/2021

- Data Hub concept creation and architecture on Azure Cloud
- Implementation of data pipelines for the ingestion in Azure Data Factory.
- Implementation of Azure Functions to pre-process ingested data
- Secure orchestration and movement of data from ADF to Azure Datalake using Azure Key Vault permissions.
- Optimal p partitioning of Datalake data in Databrick Delta Lake format
- Creations of multiple data Apache spark data aggregation scripts in PySpark and Scala
- Setup of Databrick cluster, computation of business KPIs and expose results as rest interface to Qlik, PowerBI visualization

- Setup Azure Active Directory tenant, enable Databricks AD Passthrough, Azure Apps and CosmosDB management.

Apache Spark, Big Data, Microsoft Azure

4/2017 – 1/2019

-Architecture and building of a Machine Learning Python API that is seamlessly integrated to the Spring Boot Java based platform. The Python API uses of Flask/Gunicorn server.
- Build machine learning smart alert system which is based on a combination of self-made outlier detection algorithm and publicly available outlier detection algorithms
- Build of a machine learning system for bank data prediction.
- Using XGBoost and an ensemble of other machine learning techniques for bank data forecasting.
- Deep reinforcement learning for banking network optimization. Usage of both convolution and recurrent neural network to teach multiple agents.

Big Data, Data Mining, Data Science

6/2016 – 7/2019

- Build on-premise Data Centre

- Install on-premise clusters with numerous head and worker nodes using Linux Ubuntu

- Build data Ingestion pipelines to ingest data from file protocols (FTP/SFTP), relational databases (e.g., MySQL, Postgres, Oracle, MSQL Server), document database (e.g., MongoDB), ERP systems (e.g., GreenPlum, SAP), file systems

- Architecture and building of a Data Prep Layer whereby imported data are aggregated using custom Spark function in Scala, SparkSQL, Spark UDF.

- Utilization of Hive and HiveContext to improve the speed of spark data aggregation.

- Performance tuning and optimal data structure setup of MongoDB to deliver KPI results with billions of data points within microsecond

- Build micro-services to access business KPI results from MongoDB and other internal databases.

- Build Java and Spring Boot web-based application

Apache Hadoop, Apache Spark, Apache Tomcat, Data Mining, Data Science, Java (allg.), Spring