Join Amazon's Frontier AI & Robotics team at the forefront of the next robotics revolution. As a Software Development Engineer embedded in the science team, you'll be crucial in transforming cutting-edge research into production-ready systems. The role focuses on optimizing large-scale transformer architectures for robotics applications, requiring expertise in CUDA and TensorRT to achieve maximum inference efficiency at Amazon's scale.
The position offers a unique opportunity to work at the intersection of large language models and real-world robotics applications. You'll collaborate directly with scientists, balancing deep technical optimization work with strategic input on model architecture decisions. The team is building the future of intelligent robotics through frontier foundation models and end-to-end learned systems.
At Frontier AI & Robotics, you'll be part of a team tackling some of the most challenging problems in AI and robotics, from sophisticated perception systems to adaptive manipulation strategies. The team leverages Amazon's massive computational infrastructure and real-world datasets to train and deploy state-of-the-art foundation models.
The role combines ambitious research vision with practical impact, as your work will scale to meet the demands of Amazon's global operations. You'll use NVIDIA's acceleration stack and other compilation techniques to meet ambitious performance targets. Daily activities include optimizing transformer blocks, debugging performance bottlenecks, and designing monitoring systems for production deployment.
The compensation is competitive, ranging from $129,300 to $223,600 based on location, with additional benefits including equity, sign-on payments, and comprehensive medical and financial benefits. This is an excellent opportunity for those passionate about pushing the boundaries of robotics while working with world-class researchers and seeing their innovations deployed at unprecedented scale.