We are utilizing a city's existing fleet and a growing number of mobile devices connected to our network, we collect real-time data that supports the enforcement of traffic laws. We are implementing the first technology provider to create a vision-based solution capable of detecting, understanding, and determine causation of traffic violations, to help cities achieve their Vision Zero safety goals.
- Responsible for developing computer vision and machine learning technologies for prototyping and deployment on cloud and embedded systems.
- Building road scene reconstruction system using computer vision and other sensors with following techniques:
- applied deep learning (object detection, semantic segmentation, depth prediction etc.)
- motion & trajectory estimation,
- scene understanding,
- visual odometry,
- sensor fusion,
- mutli-object tracking,
- camera calibration,
- parameter estimation & optimization
- lanes and road layout estimation
- Experimenting with real-world data to validate, evaluate and improve the algorithms.
- Responsible for productizing the CV/ML algorithms into product features, with support and guidance from other system engineering teams
Who we're looking for?
- Computer Science engineering degree or equivalent - Math, Physics or similarly quantitative-heavy background
- Experience with OpenCV and Pytorch, also other computer vision/deep learning libraries
- Experience in building visual data processing pipelines, including the design of necessary validation automation
- Ability to transfer prototyping code/algorithm into production grade feature of a working system (C++)
- Experience in Python, OpenCV, Dlib
- Experience with common Machine Learning methods
- Experience with Pytorch, Keras or TensorFlow is a plus
- Experience in using Deep Learning models in automotive projects (object detection, scene semantic segmentation etc.)
- 5+ years of experience ii IT projects, ideally similar position
NICE TO HAVE
- Experience in C++
- Knowledge of Nvidia TensorRT, DeepStream SDK
- Experience in deep learning models optimization (pruning, quantization, knowledge distillation etc.)