Software Engineer
2018 - 2021Hella Aglaia Mobile Vision GmbH- Researched and developed 3D object detection methods with deep learning for LiDAR point clouds.
- Worked on a multi-camera and LiDAR fusion approach for object detection from prototype to production. Grew the team from one to five members.
- Supervised a master thesis student on dynamic occupancy grid mapping with particle filter using radar data. Led to a joint publication.
Technologies: Python, Deep Learning, TensorFlow, Docker, C++, PCL, OpenCV, Object Detection, Tracking, Occupancy Grids, LiDAR, Keras, PyTorch, Ubuntu, Windows, Data Science, SQL, Calibration, Epipolar Geometry, Scrum, Visual Studio 2010, Jupyter Notebook, Jupyter, C, Algorithms, Artificial Intelligence (AI), Image ProcessingComputer Vision Research Contractor
2015 - 2017Zappar LTD- Developed a novel pattern and detection method. Created fast, robust, and precise localization methods for marker-based augmented reality at a room scale. Helped kickstart the new ZapBox mixed reality headset product and its production.
- Developed a visual search index for research and development (R&D) purposes.
- Developed a novel pattern detection method and OCR classifier for mobile phone research and development (R&D).
Technologies: iOS, C++, Python, Deep Learning, Bundle Adjustment, Structure from Motion (SfM), 3D Reconstruction, Simultaneous Localization & Mapping (SLAM), OCR, Data Science, Agile, Scrum, Epipolar Geometry, Jupyter Notebook, Jupyter, C, Augmented Reality (AR), Mobile Vision, Algebra, Matrix Algebra, Linear Algebra, 3D Geometric Analysis, OpenGL, Graphics Processing Unit (GPU), Algorithms, Artificial Intelligence (AI), Image ProcessingBack-end Mapping Specialist
2015 - 2015Wonderbly- Analyzed 35,000 satellite images for image quality problems, including stitching quality, to guide the second product release risk assessment. After selling a million books and getting Google ventures on board, this was the company's second product.
- Developed a reactive web labeling tool for satellite images.
- Invited a team of undergraduate students to label 35,000 satellite images and obtain a week-long internship.
Technologies: Python, OpenCV, Node.js, Angular, Machine Learning, Data Science, AngularJS, Amazon Web Services (AWS), Algorithms, Image ProcessingImage Processing Engineer
2014 - 2014Skin Analytics- Pivoted the company strategy by developing a dermoscopic mole segmentation method. This was the foundation of the company's rescue funding of £500,000 after running into startup cash-flow issues.
- Completed a government research grant worth £100,000 and focused on full-body mole screening.
- Supported the computer vision team, which developed a dry-skin beauty classifier and mobile application.
Technologies: C++, Python, OpenCV, Image Segmentation, Image Classification, Image Stitching, Jupyter, Jupyter Notebook, C, Algorithms, Artificial Intelligence (AI)Master Thesis Student
2013 - 2013BMW- Developed a massively parallel particle filter and evaluated the classifier performance regarding occupancy grid mapping and static or dynamic classification.
- Researched a novel particle filter approach and facilitated better understanding and evaluation of such method inside the R&D group, which led to new publications.
- Combined the particle filter approach with Dempster-Shafer grids, which led to the publication of a new IEEE paper.
Technologies: C++, CUDA, Occupancy Grids, Particle Filter Tracking, Sensor Fusion, LiDAR, Machine Learning, Data Science, Bayesian Statistics, Bayesian Inference & Modeling, Simultaneous Localization & Mapping (SLAM), OpenGL, Visual Studio, Visual Studio 2010, PCL, Robotics, MATLAB, MATLAB Statistics & Machine Learning Toolbox, C, Graphics Processing Unit (GPU), GPU Computing, Algorithms, Artificial Intelligence (AI), Image Processing