DevOps Engineerr, Machine Leaning

Key Skills

Java Programming, Python Programming, C++, SQL, DevOps, AWS, Jenkins, Docker, Kubernetes-K8s, Continuous Integration & Continuous Delivery - CI/CD, Terraform, Linux Operating system, Bash scripting

Job Description

The DevOps Engineer will be a member of and supporting our Machine Learning for Motion Planning software development and research team. This team introduces and advances new Deep Learning techniques for our Motion Planner. We are looking for an experienced DevOps engineer to help build and scale machine learning infrastructure, with the eventual goal of partnering with our research scientists to train models and write production level software.

Role responsibilities:

• Work with Python including Python package management, scripting, and reviewing code
• Create and deploy docker images dedicated for machine learning training
• Develop tools for deployment and monitoring of highly available, mission critical services
• Troubleshoot critical build and development issues
• Build and support cloud infrastructure including Kubernetes
• Establish and maintain a continuous integration infrastructure
• Manage/Upgrade the applications as necessary.

Impact:

• Work closely with robotics research and software engineers specialized in Machine Learning, Data, Motion Planning and Controls to identify places where data-driven deep learning approach is beneficial.

Experience to bring:

• DevOps
• Python, Bash scripting
• Cloud (AWS)
• Linux, version control, CI/CD (Jenkins)
• Terraform, Docker, Kubernetes

We'd be pleasantly surprised to see:

• Experience managing AWS and in-house server infrastructure using IaC
• Large-scale compute and build infrastructure.
• C++
• Bazel

Machine Learning at Motional:

We have produced groundbreaking advancements in the autonomous vehicle industry including

• nuScenes: large scale public data set for autonomous driving (https://www.nuscenes.org)
• PointPillars: Fast Encoders for Object Detection from Point Clouds (https://arxiv.org/abs/1812.05784)
• PointPainting: PointPainting: Sequential Fusion for 3D Object Detection (http://arxiv.org/abs/1911.10150)
• nuPlan: A closed-loop ML-based planning benchmark for autonomous vehicles (https://arxiv.org/pdf/2106.11810.pdf)
• nuPlan (continued): (https://motional.com/news/nuplan)
• Predicting the future in real time for safer autonomous driving (https://motional.com/news/technically-speaking-predicting-the-future)
• Continuous Learning Framework (https://motional.com/news/technically-speaking-learning-with-every-mile-driven)
• Mining for scenarios to help better train our AVs (https://motional.com/news/technically-speaking-scenario-mining)