Konstantin Dobrokhodov
Bis 2020, Computer Vision Engineer, State Research Institute of Aviation Systems (GosNIIAS)
SBER, Russische Föderation
Über mich
Experience: - Development of solutions for computer vision tasks based on classical algorithms and neural networks. Selection of metrics. - Training and optimization of neural networks and their output to production (detection, classification, text recognition, segmentation) - Preparation / cleaning of data, creation of instructions for marking and their control
Werdegang
Berufserfahrung von Konstantin Dobrokhodov
Bis heute 3 Jahre und 10 Monate, seit Aug. 2020
Data Scientist (Computer Vision)
SBER
Task: Development of OCR legal documents to extract text and data from stamps. Executed: - Optimization of old developments. The execution speed of classic CV algorithms has been increased by 40%, neural networks by 70%; - Code refactoring. Creation of a single format for all models; - Creation of instructions for marking, and approaches for their control; - Optimization of neural networks (ONNX, OpenVINO); - Preparation of a service for output to prod.
1 Jahr und 1 Monat, Aug. 2019 - Aug. 2020
Computer Vision Engineer
State Research Institute of Aviation Systems (GosNIIAS)
Task: Combining multi-spectral images to improve visibility for pilots. Executed: - Choice of metrics; - Image preprocessing using classical CV algorithms; - Searching and cleaning new data, creating queries to generate new data using computer graphics; - Training of neural networks, comparison of quality with alternative approaches.
Ausbildung von Konstantin Dobrokhodov
2 Jahre, Sep. 2018 - Aug. 2020
Optical engineering
Bauman Moscow State Technical University
Development of optical systems and image processing algorithms
4 Jahre, Sep. 2014 - Aug. 2018
Optical engineering
Bauman Moscow State Technical University
Development of optical systems
Sprachen
Russisch
Muttersprache
Englisch
Fließend
Deutsch
Grundlagen