Data and Cloud Engineer
Halbauer GmbH
(Automotive and vehicle construction)
Skills
Cloud architect, Senior Cloud Engineer, Senior Data Engineer, Team Lead, ML DevOps, Graphdatenbank, Kubernetes, Java, Python, NET, DevOps, test driven development, Confluence, Jira, Azure, AWS, Lambda, Azure Pipelines, AWS EMR, AWS Glue, Azure Data Factory, Azure Machine Learning, AKS / EKS, ADLS Gen2 / AWS S3, Data(base), SQL, Neo4j, Cosmos DB, Databricks / Spark, Kafka, DVC, SSDT, Delta Lake, Cassandra, .NET 6, Scala, Orchestration, versioning and tracking, Terraform, Kubernetes, Docker, Ansible, MLFlow, Argo, Luigi, Airflow, git, Github, Github Actions, Visualization, Power BI, Grafana, Machine learning, algorithm, Regressions, SVM, ensemble methods, XGboost, RM, clustering, KNN, naïve Bayes, graphic models, Deep learning, Neural Network, NLP, Autoencoder, Bayesian, MCMC, correspondence analysis, Spark, data ingestion, data visualization, Databricks, algorithms, DB, Azure Functions, AKS, big data, DWH, batch processing, Hadoop, backend, Excel, Node.js, R, Shiny, GCP, Azure, AWS
Project history
09/2021
-
07/2022
In order to migrate a classical machine learning model to Azure, which aims to predict the vehicle repair rate, I interpreted the Python project into Spark and used Databricks, MLflow and Azure Data Factory to automate the data ingestion and to parallelize the model training. Front-end data visualization uses Grafana. Tools: Azure Data Factory, Databricks, Spark, Delta Lake, Azure Pipelines, Grafana.
02/2020
-
08/2021
Data Scientist and Cloud Engineer
Island Labs GmbH
(Architecture and civil engineering, 10-50 employees)
Kitchen furniture blueprints sometimes contain errors which should be avoided before being sent to construction. As a
technical team working with a kitchen provider, we received thousands of kitchen plans on a daily basis. I designed a graph model to represent the furniture of a kitchen, geometric and graph algorithms to learn and look for errors in them, and a pipelined workflow on Azure to process the data. Kitchen blueprints as data are first fed into the workflow by Kafka in real time, these are then validated and sent to a Neo4j cluster by an Azure function. After the kitchen blueprints get ingested and analyzed by Neo4j, the result is sent to the Cosmos DB and returned to the client, in case an error is detected or a warning occurred. Tools: knowledge graph, Kafka, Neo4j, Azure Functions, AKS, Python and .NET.
technical team working with a kitchen provider, we received thousands of kitchen plans on a daily basis. I designed a graph model to represent the furniture of a kitchen, geometric and graph algorithms to learn and look for errors in them, and a pipelined workflow on Azure to process the data. Kitchen blueprints as data are first fed into the workflow by Kafka in real time, these are then validated and sent to a Neo4j cluster by an Azure function. After the kitchen blueprints get ingested and analyzed by Neo4j, the result is sent to the Cosmos DB and returned to the client, in case an error is detected or a warning occurred. Tools: knowledge graph, Kafka, Neo4j, Azure Functions, AKS, Python and .NET.
10/2018
-
01/2020
Data scientist and engineer
(Public service)
Astronomical data can be immense and requires a big data solution to process it. Working together with astronomers, I
designed a Spark application for an on premise DWH to process the data and read in from and write out to its different layers. The application includes both batch processing to deal with existent data and streaming to deal with new data coming in. I also optimized the Spark application at different levels which has improved its performance remarkably. Tools: Spark 2.x, Kafka, Hadoop, Java.
designed a Spark application for an on premise DWH to process the data and read in from and write out to its different layers. The application includes both batch processing to deal with existent data and streaming to deal with new data coming in. I also optimized the Spark application at different levels which has improved its performance remarkably. Tools: Spark 2.x, Kafka, Hadoop, Java.
04/2018
-
09/2018
Biostatistics and AI Developer
Firalis SA
(Pharmaceuticals and medical technology, 50-250 employees)
The combination of omic data and machine learning / deep learning is one of the cutting-edge areas of biopharma research. In this project where the company aims to get a better understanding of the high-throughput RNA sequencing data of patients undergoing different medical treatments after a myocardial infarction, my machine learning and deep learning algorithms came to help. Tools: Python and R.
07/2017
-
08/2017
Psychometry and AI Developer
Centre Hospitalier de Rouffach
(Public service, 50-250 employees)
In the psychomotor therapy the diagnosis of schizophrenia had been to that date carried out by the observation and the subjective judgment of medical professionals. With the hope to develop an AI assistant helping physicians make better and more accurate clinical decisions based on more quantitative indicators, I built an application using machine learning and statistical models to predict schizophrenia. Tools: R.
02/2016
-
08/2016
AI backend developer
AI-Developer
(Pharmaceuticals and medical technology)
As a backend developer I participated in the development of a diagnosis assistant system, the goal of which is to detect among others pulmonary sarcoidosis using AI. Tools: .NET.
11/2015
-
02/2016
Risk analyst
Bank of Communications
(Banks and financial services, 5000-10.000 employees)
Functions: As a risk analyst I ensured the role of developing statistical tools to monitor credit risk, market risk, liquidity risk and operational risk of the bank. Tools: R and Excel.
Local Availability
Open to travel worldwide
Bis zu 100%
