Machine Learning Engineer - Vision

Play a part in shaping the future of human communication technology. Contribute to a unique multidisciplinary system that models and understands human interaction, redefining what's possible with computer vision, physics, and signal processing at the edge.

Description
In this role, you'll be at the forefront of a one-of-a-kind technical challenge, developing and implementing novel methods for processing and enhancing a proprietary sensor that models human communication. You'll work hands-on with exclusive data: designing the algorithms that process it, and defining the metrics that evaluate and drive its continuous improvement.

Your insights will carry real weight, directly informing sensor decisions and shaping the architecture of the broader system. You'll lead multi-level research efforts to advance a truly unique sensor, drawing on a rich and diverse technical toolkit spanning signal processing, computer vision, physics, and state-of-the-art deep learning. You'll own proprietary data collections using high-end computer vision techniques, studying signals from their raw-level behavior all the way through to their top-level impact on product performance.

This is a role that lives at the intersection of deep research and real-world impact. You'll conduct cutting-edge investigations and translate your findings directly into product outcomes, influencing decisions across the full stack, from hardware choices and algorithmic pipelines to the features that reach the final product.","responsibilities":"Develop and implement novel algorithms for modeling and understanding human communication, combining 2D/3D computer vision, signal processing, and deep learning
Work with unique proprietary datasets - design large-scale data processing pipelines, define quality metrics, and provide actionable feedback to improve data collection and labeling workflows
Devise and implement rigorous evaluation frameworks to measure model and data quality, and drive continuous improvement across the system
Design neural network architectures optimized for SOTA accuracy and computational efficiency
Stay current with the latest research across computer vision, signal processing, and efficient ML; evaluate and integrate relevant advances into the team's work
Contribute to internal tooling and best practices for reproducible, scalable ML research and deployment

Preferred Qualifications
Knowledge and experience with 3D data (e.g., point clouds, depth sensing, 3D reconstruction)
Ph.D. in a relevant field
Hands-on experience with model compression techniques (quantization, pruning, distillation)
Experience with MLOps workflows and CI/CD pipelines for model development

Minimum Qualifications
M.Sc. in Computer Science, Electrical Engineering, or a related field, with a thesis in AI, computer vision, data science, or an equivalent discipline
At least 3 years of hands-on experience in machine learning
At least 3 years of hands-on experience in image processing and computer vision
Strong foundation in deep learning theory and practical experience training large-scale models
Proficiency in Python and deep learning frameworks such as PyTorch
Background in signal processing and physics-based modeling
Practical experience with large-scale data processing, pipeline design, and performance evaluation
Experience utilizing modern frameworks and keeping up with recent research