I am a versatile software & ML engineer with expertise in backend development, machine learning, cloud infrastructure, and data pipelines. I consistently deliver high-quality solutions with a combination of technical expertise, creative problem-solving, and excellent collaboration skills.
Python
Django
Django Rest
Sendgrid
Celery
Docker
PyCharm
C#
Nginx
Sentry
Github
Gitlab
Jira
Grasshopper
Linux
Kanban
Highly skilled in Python. Created and maintained RESTful APIs in Django & Django REST, as well as web applications such as www.airteam.cloud. Experienced in containerization with Docker.
AWS
GCP
S3
EC2
Load Balancing
Big Query
PubSub
Beam
Spark
PostgreSQL
Gitlab CI CD
Grafana
Terraform
Experienced in working on multiple cloud platforms with infrastructure as code and via UI. Built CI/CD pipelines with custom gitlab runners for automatic testing & deployment. Managed various web applications on AWS and GCP. Built ETL pipelines with Apache Beam running with DataFlow.
PyTorch
MlFlow
Tensorboard
Tracking
Pointclouds
Images
Object Detection
Experienced in different deep learning fields such as Object Detection & Tracking, Image & Pointcloud segmentation, as well as general classification problems. Skilled in setting up and using MlFlow for ML Lifecycle & experiment management, and model serving.
Research in deep learning based multi-object tracking
I worked with Max in the context of a large image processing project at the Deutsche Bahn. In the project we were investigating the quality of various state-of-the-Art anomaly detection algorithms for a fixed camera system problem. While working with Max I found him to be a very knowledgeable, curious and hard working. He was able to explain his work and process very clear both verbally and in written form. His work process was very clean and structured. On a personal level, it was a pleasure to work with Max as he quickly integrated into our team and made sizable contributions. I would highly recommend Max as a developer and believe he would also excell as a lead. He would be an asset to any team.
For about six months, Max worked on his Master's thesis in our team. His work ethic and time management skills impressed me as his supervisor. He always went above and beyond what we expected to achieve. He was always willing to put in the time and effort to do the research and find professional and scientific answers to the open questions during his research. Any team would be lucky to have Max on board and I would be delighted to recommend him to any company in need of new talent.
We describe a new approach for multi-pedestrian tracking in geo-referenced aerial imagery called AerialMPTNet, which fuses appearance features, movement predictions, and pedestrian interconnections. We also introduce the Aerial Multi-Pedestrian Tracking (AerialMPT) dataset, which we believe to be the largest and most diverse dataset to date, and evaluate AerialMPTNet on AerialMPT and KIT AIS, showing significant improvement over other methods in accuracy and time-efficiency.
We present a solution for improving identity recognition in a car cabin using Time Series Classification (TSC) and deep neural networks. The input data, gas and brake pedal pressure, is easily collected during driving and the combination of LSTM and ResNet classifiers leads to an accuracy of 79.49% on a 10-drivers subset of NUDrive and 96.90% on a 5-drivers subset of UTDrive, outperforming other models by a large margin.