Research Scientist

pony.ai - Fremont, CA (30+ days ago)


Company Mission
Our mission is to revolutionize the future of transportation by building the safest and most reliable technology for autonomous vehicles. Armed with the latest breakthroughs in artificial intelligence, we aim to deliver our technology at a global scale. We believe our work has the potential to transform lives and industries for the better.

Research Arm
The research arm will focus on core problems in machine learning for autonomous driving.

Computer vision: we need to solve important computer vision problems such as sensor-fusion 3D object detection, traffic light detection, emergency vehicle detection, activity recognition, scene understanding.
Prediction: we need to predict where actors will go. This is very challenging as human behavior is inherently unpredictable.
Planning: our software needs to drive like a human. It needs to infer human actors’ behavior and respond to them in complex situations such as merging and lane changes. It needs to be super human in terms of safety and reliability.
Localization and mapping: HD maps make several core problems such as perception and planning easier. To be scalable, it is essential we automate the map building as much as possible. We also need to reliable localize the autonomous vehicle in both urban and highway settings.
Deep learning and deep reinforcement learning: deep learning and deep reinforcement learning hold the promise to solve many hard edge cases. However, there are many fundamental problems unsolved. For example, the representations learning by deep networks is neither robust (vulnerable to input perturbation) nor abstract (e.g. does not generalize to novel viewpoints). There has not been a lot of progress in multi-agent deep reinforcement learning

Job Description
As a research scientist, you will collaborate with our current team working on deep learning, computer vision, and autonomous driving. You will also engage with the international research community by publishing research papers and attend conferences. We value high-impact technical solutions that advances the state-of-the-art. Please join us If you are passionate about autonomous driving, have publications at top conferences or journals in the field of computer vision, machine learning, robotics and are an excellent programmer.

Requirements

MS/PhD degree in a machine learning related discipline or equivalent practical experience
Experience in one or more of the following: computer vision, sensor fusion, object tracking, deep learning and deep reinforcement learning
Experience in data structures and algorithms is a plus
Strong programming skills in C++ and/or python, experience in Tensorflow or PyTorch