Géza Velkey Email & Phone Number

Stuttgart, Baden-Württemberg, Germany

Menlo Park, California, United States

Technical lead on multiple unsupervised learning projects.

Menlo Park, California, United States

Research engineer responsible for unsupervised and self-supervised learning projects. Covered full machine learning lifecycle from model research to integration to embedded systems.

Budapest, Hungary

Machine Learning specialist and cloud architect responsible for development and deployment of cutting-edge internal ML models.Research and train sparse 3D neural networks for semantics understanding in urban scenarios. Replaced rule-based ecosystem with fully ML driven multi-GPU implementation. Deployed ML ideas at scale on different cloud environments in both public, private and Azure government cloud offerings. Created large scale cloud-native ML ecosystem from scratch with full pipeline automation and deployment process on Kubernetes.

Stuttgart, Baden-Württemberg, Germany

Azure cloud architect, applied deep learning in the field of advanced text processing and semantic understanding. Product owner on ALM product.Automated cloud deployment processes with Python to Kubernetes using Helm and Azure cloud Python APIs. Reached full automation for instance provisioning, monitoring and maintenance. Built up Swagger APIs in a self-developed microservice architecture suitable for large-scale ML inference and running the SaaS product in the same cluster.Released several public applications on MS marketplace, built up Azure architecture to support it.

Budapest, Stuttgart, Sunnyvale

Product owner on level 5 automated driving AI research. Applied state of the art deep learning models to solve highly relevant topics for automated driving.Responsible for development and deployment infrastructure, integrated complete Docker ecosystem for training neural networks in the cloud, local clusters and on the edge.Implemented a complete training and inference framework for neural networks on clusters and running on the edge in automated driving with support for PyTorch and Tensorflow.Coordinates the work of a scrum team on the research project using modern agile methodologies.

Renningen, Baden-Württemberg, Germany

Budapest

Applied machine learning research to solve level 4-5 automated driving problems. Deep learning helped to advance in technology in a very quick iterative manner.Worked with devops and model development in a daily manner. Interface implementation was also an important part of the work.

Budapest

Navigation development in C++. Implemented numerous features for head units in production, learned and applied new cpp features introduced with c++11. Participated in regular code review and learned advanced build ecosystems and test automation.

Budapest

On the Magyar Telekom's One.ERP project involved in:- Testing of SAP middleware systems and interfaces- Oracle Service Bus middleware deployments- Creating automated test environment in JAVA for SOAP web services (OSB, SOABP)- Developing automated documentation creation scripts- Server infrastructure sizing analysis based on predicted load produced by SAP messages- Microsoft Office scripting in Python and VBA- Automated architecture diagram drawing scripts using yEd,JAVA and VBA- Command line programs in Windows- SAP extraction tool creation in Python- Usage of file-based systems like SOABP EFB, IBM MQ

Master Of Science - Computer Science, Artificial Intelligence, Highest Honors

Master of science in computer science - machine learning major with processor optimization topics.

Bachelor’S Degree, Electrical And Electronics Engineering, Highest Honors

Activities and Societies: International Physics Olimpiad BSc in electrical engineering, major in AI research and applied deep learning.

Bachelor’S Degree, Mathematics

Studied applied Mathematics BSc along with electrical engineering to get a better understanding of the underlying mathematical theories.