Bringing self-driving vehicles to our roads is the most transformative opportunity of our generation. Aurora is taking a fresh start with the development of self-driving technology, combining excellence in AI, rigorous engineering, and a team with decades of experience building robots that work.
Led by a team of seasoned experts, including three of the world's leaders of self-driving technology, our mission is to deliver the benefits of self-driving technology safely, quickly, and broadly. We are designing the software and hardware to power the transportation of our future that will make our roads safer, give more people access to mobility, and reduce congestion and pollution in cities - improving the quality of life for all. The challenge in what we are endeavoring to achieve is transcendent; we are developing perhaps the world's most complex computing system and asking it to perform the task of transporting and keeping safe our most precious asset: human life.
We're looking for people who are as excited as we are to solve these complex problems and make this tremendous impact on our future, and who want to be surrounded by great people while we do it. We are searching for a Deep Learning Inference Engineer to work closely with machine learning experts to productionize our state-of-the-art networks to run as fast as possible on our self-driving cars.
* Work with machine learning experts to design and implement networks optimized for GPU-accelerated inference
* Develop advanced inference engine techniques to enable real-time deployment of models
* Deploy productized networks to run efficiently on our hardware
* Develop tools for profiling, analyzing, and improving both training and inference performance
* BS/MS/ or PhD in Computer Science or a related field / equivalent experience
* Intimate understanding of at least one open-source deep learning framework (e.g., Tensorflow, Caffe, Torch)
* Excellent C++ programming and software design skills
* Proven GPU programming (e.g., CUDA, OpenCL) track record
* Comfort with vector processing techniques
* Code contributions to an open-source deep learning community
* Authored custom operators down through the lowest levels
* Experience using GPU-accelerated libraries (e.g., cuDNN and cuBLAS)
* Experience with code generation & optimization (e.g., Halide)
* Experience with network model compression and quantization
Working at Aurora
Our work has real purpose. Delivering self-driving will improve lives around the world, expanding access to transportation, revitalizing cities, giving people more time back every day.
We're one team. We're inspired by the challenge of what we're solving and the impact our work will have on society. Our camaraderie is built on respect for our work and the fundamental belief our success will be a result of working together.
The Founding Team
Aurora has assembled the most experienced leadership team in the space. Chris Urmson helped lead Carnegie Mellon's efforts in Darpa's Grand Challenges, then was a founding member of Google's self-driving team. Sterling Anderson worked on the tech at MIT before leading Tesla's Autopilot system. Drew Bagnell, also a Carnegie Mellon alum, is a machine learning expert who helped build Uber's autonomy effort. At Aurora, these three continue to bring experts from all areas of the industry to the team. We are funded by some of Silicon Valley's best venture capital firms, including Greylock and Index Ventures.
Aurora works at the intersection of rigorous engineering and applied machine learning to transform the way people and goods move.