Upgrade Post project
Profileimage by Anonymous profile, Data Scientist, Software Achitect/Engineer, AI Specialist
Bad Homburg
not available until 12/31/2024

Last update: 24.12.2023

Data Scientist, Software Achitect/Engineer, AI Specialist

Graduation: Mathematics
Hourly-/Daily rates: show
Languages: German (Native or Bilingual) | English (Full Professional) | Russian (Native or Bilingual)

Keywords

Skills

Machine Learning, Data Science, Algorithms, Data Structures, Programming, Object-oriented design, DevOps, Java, Python, C#, NET, JavaScript, TypeScript, Scala, Tensorflow, Keras, Pandas, Numpy, Jupyter Notebook, IntelliJ, PyCharm, Docker, Microservices, Jenkins, Unity, Git, GitLab, GitHub, SVN, SQL, NoSQL, Firebase, MongoDB, Cassandra, Hibernate, JPA, Kafka, CSS, LESS, HTML, JSON, XML, REST, SSH, Linux, Maven, Angular2, Vue, Visual Studio, J2EE, VisualVM, ETL, Big Data, Kubernetes, AWS Cloud, Prometheus, Grafana, Postgres, database, algorithm, Reinforcement Learning, Deep Learning, SCRUM, Game Engine, Continuous Testing, web service, image recognition, REST web service, Flask, Heuristic, Oracle SQL DB, Spring, JAX-RS, JAXB, functional specifications, REST Web Services, Kubernetes (K8s), Helm

Project history

06/2019 - Present
Big Data Engineer @ Deutsche Bahn Passenger Information
Deutsche Bahn (Transport and Logistics, >10.000 employees)

Delivering passenger information to Deutsche Bahn customers, such as arrival/departure times, delays or outages,  requires live tracking of all train movements in Germany. This means the IT system has to be able to process millions of events per day.  This is achieved by a modern event-driven, microservice architecture. As a big data engineer I am responsible for the ongoing development of this application.

RESPONSIBILITIES AND TASKS
  • Design, development and operation of microservices
  • Event-driven Architecture based on Kafka
  • Kubernetes based infrastructure on AWS  Cloud
  • Helm driven deployment
  • Extensive monitoring of the production environment with Prometheus and Grafana.
  • Postgres database
08/2019 - 01/2020
Success Rate Estimation @ Conver GmbH
Conver GmbH (Marketing, PR and Design, < 10 employees)
Direct Marketers can increase their margins by allocating more
marketing resources on customers that tend to order more products.
In collaboration with Conver GmbH we developed a Machine
Learning pipeline to generate such knowledge and extended the
business processes of Conver GmbH to apply this knowledge.

RESPONSIBILITIES AND TASKS

* Business Analysis
* Machine Learning consulting
* Development of an ETL pipeline based on python, pandas,
jupyter notebook, ...
* Development of a Machine Learning model to estimate
success rates.
06/2018 - 05/2019
AI based package classification
KION/Dematic (Transport and Logistics, >10.000 employees)
Automatic handling of products requires an extensive knowledge about  the shape and form of the packaging. Up until now this data was collected in an entirely manual process which is highly error-prone. I proposed, prototyped and developed a classification service that automates this process. The software improves on the classification accuracy of humans and therby increases the performance of the industrial plants.
09/2016 - 05/2019
Dematic Multishuttle optimal control algorithm
KION/Dematic (Transport and Logistics, >10.000 employees)
The Dematic Multishuttle system (https://goo.gl/PQFACb) is a high throughput buffer storage for small packages and parts. I proposed, implemented and supervised the development of a new control algorithm based on Machine Learning techniques and heuristic optimization. The algorithm has to schedule, sequence and route thousands of package movements  every hour (agent decisions approx. every 50ms).
09/2013 - 09/2016
3D-Bin-Packing
KION/Dematic (Transport and Logistics, >10.000 employees)
Dematic AMCAP is a robot system that is able to build heterogeneous pallets on its own. To perform this task one has to create a plan of the pallet (compute positions of each package) that the robot system can follow. Planning stable, shop-friendly pallets is a highly complex optimization problem (also known as 3D-Bin-Packing in Mathematics and Computer Science), which is evident by the fact that only a couple teams could master this problem. In just about a year our team has created a product that outperforms programs that were developed for multiple decades. https://goo.gl/ACqviY

Local Availability

Open to travel worldwide
100% vor Ort Einsatz im Rhein Main Gebiet
40% vor Ort Einsatz sonst.
Profileimage by Anonymous profile, Data Scientist, Software Achitect/Engineer, AI Specialist Data Scientist, Software Achitect/Engineer, AI Specialist
Register