Perception / Computer Vision Software Engineer

Key Skills

CUDA/GPU programming, Deep Learning, GPUs, Neural Networks, PyTorch, Ray.io, Spark Core, SQL, TensorRT

Job Description

Figure is an AI Robotics company developing a general purpose humanoid. Our Humanoid is designed for corporate tasks targeting labor shortages and jobs that are undesirable or unsafe. We are based in Sunnyvale, CA and require 5 days/week in-office collaboration.

Figure’s vision is to deploy autonomous humanoids at a global scale. Our AI team is looking for Localization and Mapping Software Engineers to empower Figure humanoid robots to perform highly dynamic operations in demanding real-world environments.

Responsibilities:

• Develop novel computer vision formulations, models, and algorithms for a wide variety of tasks
• Own the full development life cycle from data preparation to model deployment
• Perform training of multi-task deep neural networks
• Design and build data and evaluation pipelines
• Engineer and ship high quality, reliable software that will be deployed to the real world
• Collaborate with other Figure team members to develop and continuously improve the full autonomy stack

Requirements:

• Extensive experience in production-level software & ML engineering and best practices
• “White-box” understanding of deep learning: model details, loss formulations, and training algorithms
• Strong mathematical fundamentals in probability, optimization, and linear algebra
• Domain expertise in at least one of the following areas: 3D reconstruction, depth estimation, neural rendering, object detection, tracking & state estimation, instance and semantic segmentation
• Comfortable with fundamental geometric concepts in computer vision, such as perspective transformation, pinhole camera model, and epipolar geometry
• Extensive experience with one of the major deep learning frameworks, preferrably PyTorch
• Capable of quickly writing massive amounts of high quality, well-tested, perception software
• Thrive in a high pace environment, where solutions are often unclear and require exploration
• Passionate about building humanoid robots

Bonus Qualifications:

• Experience with CUDA and GPU programming
• Experience with scalable data engineering toolchains such as Spark,Ray, SQL
• Experience in neural network deployment toolchain and techniques, such as TensorRT and model quantization
• Publication record of developing and deploying perception solutions