Matthias Wisniowski
Verified Expert in Engineering
Computer Vision Developer
Vienna, Austria
Toptal member since December 30, 2022
Matthias is a full-stack computer vision engineer. He covers everything from camera hardware selection and calibration, data collection, labeling, curation, and modeling to building scalable experimentation platforms. Matthias enjoys solving problems holistically and learning everything necessary to impact the bottom line.
Portfolio
Experience
- Python - 10 years
- Data Science - 8 years
- Computer Vision - 8 years
- Machine Learning - 8 years
- OpenCV - 8 years
- Robotics - 4 years
- PyTorch - 2 years
- Amazon Web Services (AWS) - 1 year
Availability
Preferred Environment
MacOS, Vim Text Editor, Amazon SageMaker, Jupyter Notebook, Ubuntu, Python, PyTorch, OpenCV
The most amazing...
...project I've completed was a traffic light detection model for self driving cars allowing a fleet of 200 cars to safely navigate the streets of San Francisco.
Work Experience
Developer
Freelance
- Developed a machine learning (ML) platform for image object detection models. Used Amazon SageMaker for training and processing jobs and AWS Step Functions for workflow orchestration.
- Built an internal full-stack app using Python and React for image dataset curation and annotation.
- Deployed a Kubernetes cluster and ML platform on-premises.
Senior Data Scientist
craftworks
- Led computer vision projects in wood processing, agriculture, and manufacturing.
- Worked on 3D camera sensor selection. Developed a deep learning model (instance segmentation), sensor fusion, and multi-object tracking.
- Developed the internal TÜV TrustedAI certification process.
- Became the technical architect for the data science team.
Engineering Manager
Cruise
- Received nine patents for self-driving cars in the United States.
- Co-founded Cruise's internal active learning platform, which uses cloud-based machine learning to find the following valuable data to label. Evolved the datasets used to train car models continually.
- Managed a team of five engineers. Led the automation of human labeling for sensor data, including camera, LiDAR, and radar. Developed novel sensor-fusion algorithms and object detection methods leveraging humans in the loop.
- Built a detection framework for traffic lights that allowed the self-driving car to navigate 99.9% of all intersections safely. Included a deep learning model for the framework and a sensor fusion algorithm for probabilistic state inference.
- Developed the first computer vision stack for the level four self-driving car platform. Included sensor selection, on-vehicle placement/coverage decisions, calibration, and a GPU-based image processing pipeline.
- Built an instance detection model to aid human annotation of HD maps. Resulted in a reduction of human map annotation time from three months to ten minutes for the entire city of San Francisco.
- Took the lead for a business-critical project in another department. Led a team to rewrite the routing algorithm, making turn-by-turn decisions for the self-driving car. Invented a novel probabilistic, constraint-based routing algorithm.
Experience
Active Learning Platform for a Computer Vision Research Project in Agriculture
Education
Master's Degree in Robotics, Cognition, Intelligence
Technical University of Munich - Munich, Germany
Bachelor's Degree in Medical Computer Science
Technical University Vienna - Vienna, Austria
Skills
Libraries/APIs
OpenCV, PyTorch, React, TensorFlow, Eigen, Keras
Tools
Vim Text Editor, Git, Amazon SageMaker, Jira, AWS Cloud Development Kit (CDK)
Languages
Python, SQL, TypeScript, C++, Java
Platforms
MacOS, Jupyter Notebook, Ubuntu, Docker, Amazon Web Services (AWS), Kubernetes
Frameworks
Spark
Storage
PostGIS, Google Bigtable, Google Cloud
Other
Robotics, Computer Vision, Machine Learning, Data Science, Artificial Intelligence (AI), Image Recognition, Random Forests, APIs, Back-end, Point Cloud Data, Deep Learning, Data Visualization, Data Wrangling, Software Development, Biomedical Skills, Robot Operating System (ROS), TimescaleDB
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring