This job is no longer available
There are still lots of open positions. Let's find the one that's right for you.
The position involves research on artificial intelligence (AI) methods, specifically focusing on deep learning algorithms to enhance the performance of Simultaneous Localization And Mapping (SLAM) systems. The role requires designing algorithms in C++ that integrate SLAM with lane detection vision to improve SLAM results. Additionally, the candidate will train a deep neural network using Python and TensorFlow for SLAM based on SuperPoint, aiming to achieve significantly better results than the traditional SIFT (Scale-invariant feature transform). The position also involves extrapolating the design of the last 25 layers of the deep neural network and implementing them as post-processing in C++ to enable multithreading on multi-DSP (Digital Signal Processor). Furthermore, the candidate will build a traffic light recognition API for both x86 and aarch64 computer architectures, facilitating computer vision recognition of traffic lights from image and video input files. The integration of this API into the Advanced Driver-Assistance System (ADAS) API is also a key responsibility. Lastly, the role includes flashing and setting up operating systems on circuit boards using serial port and ADB (Android Debug Bridge).
