About the company

Every physical good around you sat in a warehouse, and every warehouse tracks their inventory. Almost 100% of warehouses track their inventory using handheld barcode scanners or even paper and pencil, and people climbing scissor lifts. It's dangerous, slow, and full of human error, resulting in $30B of missing inventory annually... from inside their own facilities!

We're Corvus Robotics. We invented Corvus One, the world's first warehouse inventory drones with Level 5 Autonomy®. They use AI and self-driving technology to fly up and down warehouse aisles day and night, helping companies' and governments' warehouses manage their inventory, a crucial operation every warehouse conducts. We're a mission-driven company improving the world's supply chain efficiency, reducing waste, and improving human productivity in a globally essential industry. Come help scale our warehouse inventory AI drones across America!

About the role

We are seeking passionate Computer Vision / Machine Learning Software Engineers to build efficient models for production robotics with constrained compute. You'll tackle diverse business and technical challenges, implementing and deploying models across our robot fleet. Our robots generate vast amounts of sensor data and we are constantly searching for ways to use this data to provide customers with deeper insights and to enhance our robots’ environmental awareness. We value problem-solving, innovation, and continuous learning, and we encourage exploring new technologies to advance our machine learning capabilities.

• US work authorization preferred, can sponsor visas

• Location: In-office hybrid (Mountain View, CA)

Join our team and help shape the future of robotics and machine learning!

What you'll do

• Develop solutions for advanced computer vision tasks, including:

  • Monocular and stereo depth estimation

  • Learning-based structure-from-motion

  • 3D occupancy networks

  • Scene understanding

  • Object detection

• Optimize performance, accuracy, and speed of compute-constrained models

• Collaborate with cross-functional teams to deploy CV/ML models into production

• Manage and improve ML pipelines and infrastructure for dataset management, training, and deployment

• Participate in R&D initiatives (20% of time) to experiment with state-of-the-art techniques

• Drive business value by leveraging your work across robotics, software, and deployment teams

Must have

• Adaptive and desire to assume responsibility in a fast-paced startup environment

• 5+ years of industry experience in Computer Vision / Machine Learning, Python/PyTorch

• Expertise in 2D and 3D computer vision techniques

• Proficiency with Linux, Git, AWS/GCP, and CI/CD workflows

• Experience in performance engineering for deep neural networks (both training and inference)

• Knowledge of model optimization techniques for embedded systems, including knowledge distillation, model quantization, and network pruning

Nice to have

• Experience with C/C++

• Familiarity with ROS (Robot Operating System)

• Knowledge of model deployment frameworks (e.g., TensorRT, RKNN, OpenVINO, ONNX, CoreML)

• Experience with TypeScript/JavaScript and Django

• Experience with data labelers, tooling, and data management