08/14/2023 updated
20 % partially available
Python Developer, Image Processing
Cluj-Napoca, Romania
Only remote
BSc Computer Science, Master's Degree Optimization of Computational ModelsAlgorithmsComputer VisionAutomationBusiness ProcessesC++ (Programming Language)Virtual Camera SystemsScalabilityImage ProcessingPython (Programming Language)SAP NetWeaver
▪ Skilled Software Developer with 10 years of experience in Image Processing / Computer Vision, fluent in Python and C++
▪ Designed and developed the Ground Truth System for three final driver monitoring solutions delivered to PSA, Hyundai, & BMW.
▪ Developed a new computer vision algorithm for monocular camera system to estimate in real time the 3D road surface in front of the car.
▪ Oversaw and managed the Human Annotation team with ~ 200 Head Count while managing a 1M€+ budget.
▪ Created and implemented a web-services layer for the Siemens Framework to improve system scalability and optimize development time for new features.
▪ Spearheaded a project to automate the company's business processes by developing a new SAP Netweaver application to increase system efficiency.
▪ Designed and developed the Ground Truth System for three final driver monitoring solutions delivered to PSA, Hyundai, & BMW.
▪ Developed a new computer vision algorithm for monocular camera system to estimate in real time the 3D road surface in front of the car.
▪ Oversaw and managed the Human Annotation team with ~ 200 Head Count while managing a 1M€+ budget.
▪ Created and implemented a web-services layer for the Siemens Framework to improve system scalability and optimize development time for new features.
▪ Spearheaded a project to automate the company's business processes by developing a new SAP Netweaver application to increase system efficiency.
Experience
Professional experience0 - 5 years
Freelancer since2022
Languages
GermanBasic knowledgeEnglishFluentRomanianNative speakerHungarianNative speaker
Project history
▪ Served as a technical expert for and liaison between AROBS and Continental to communicate and present research results and technical details on the performance of algorithms, data marking, tools, and other solutions.
▪ Designed and developed the Ground Truth System for three final solutions delivered to PSA, Hyundai, and BMW.
▪ Oversaw and managed the Human Annotation team with ~ 200 Head Count while managing a 1M€+ budget.
▪ Selected and managed a team of suppliers to produce high-quality ground truth data for driver monitoring, resulting in a 25% reduction in data errors and a 50% increase in data collection efficiency.
▪ Delivered a dataset with an accuracy of 99.9% (including corner cases) and a 90% reduction in false positives by implementing new validation protocols and performance metrics.
▪ Developed a new tool for validating driver monitoring algorithms that streamlined the validation process by automating repetitive tasks, reducing manual labor by 80%, improving the overall performance of algorithms.
▪ Designed and developed the Ground Truth System for three final solutions delivered to PSA, Hyundai, and BMW.
▪ Oversaw and managed the Human Annotation team with ~ 200 Head Count while managing a 1M€+ budget.
▪ Selected and managed a team of suppliers to produce high-quality ground truth data for driver monitoring, resulting in a 25% reduction in data errors and a 50% increase in data collection efficiency.
▪ Delivered a dataset with an accuracy of 99.9% (including corner cases) and a 90% reduction in false positives by implementing new validation protocols and performance metrics.
▪ Developed a new tool for validating driver monitoring algorithms that streamlined the validation process by automating repetitive tasks, reducing manual labor by 80%, improving the overall performance of algorithms.
* Built an inhouse driver monitoring algorithm which served as a base of comparison for future external contracts.
* Developed an algorithm for camera synchronization.
* Contributed to a team effort in developing a solution to adjust the camera configuration in real time, enabling fast
adaption (<1s) to changing lighting environments.
* Maintained a C++ library for multiple projects to increase code reuse and decrease development time.
* Led and managed ML projects to design and build artificial neural network systems, including a 3D gesture recognition
algorithm, improving the system's overall performance by 40%.
* Developed an algorithm for camera synchronization.
* Contributed to a team effort in developing a solution to adjust the camera configuration in real time, enabling fast
adaption (<1s) to changing lighting environments.
* Maintained a C++ library for multiple projects to increase code reuse and decrease development time.
* Led and managed ML projects to design and build artificial neural network systems, including a 3D gesture recognition
algorithm, improving the system's overall performance by 40%.
* Led the pre-development of a new embedded C++ component, for monocular camera system, to estimate the 3D road
surface model in front of the car.
* Reduced costs 30% by delivering an algorithm running on limited hardware.
* Oversaw algorithm development, boosting the accuracy of 3D reconstruction around the car using ego motion.
surface model in front of the car.
* Reduced costs 30% by delivering an algorithm running on limited hardware.
* Oversaw algorithm development, boosting the accuracy of 3D reconstruction around the car using ego motion.