Machine Learning Engineer

Radar Reinvented.

Echodyne offers the world’s first compact solid-state true beam-steering radar for a wide range of industries and applications. Our high-performance radars work in all weather and are designed for autonomous vehicles, uncrewed aircraft & drones, and security of borders, critical infrastructure, and smart cities. The company combines the patented technology of metamaterials with powerful software to create a radar sensor with unprecedented performance at commercial price points. Echodyne offers its radars to companies working in Automotive, Transportation, Critical Infrastructure Protection, Border Security, Smart Cities, Uncrewed Aircraft Systems (UAS), and Airspace Management including Urban Air Mobility (UTM).

Echodyne is seeking a Senior Machine Learning Engineer to join our fast-growing team. 

ROLE OVERVIEW

In this role, you’ll develop machine learning capabilities for Echodyne radars and data systems. You’ll work with radar data to enhance our labeling process and improve how our systems detect, classify, and track targets in complex, changing environments.

You’ll own work from data exploration and model prototyping through evaluation and production deployment. You’ll make practical tradeoffs among accuracy, latency, computation, and reliability.

Your work will have a direct impact on our advanced data systems and growing fleet of fielded radars, helping make them more robust and effective at protecting the infrastructure and people that rely on them in an evolving world.

RESPONSIBILITIES

• Prioritize and ensure real-world robustness and zero-regression deployment of models to our fleet of fielded radars.
• Improve the accuracy and capacity of our data labeling systems by developing active learning methods and automated labeling systems, and by advising on tool design.
• Design data augmentation methods that reflect realistic variation in radar deployment environments.
• Improve existing track-classification models through hyper-parameter tuning and trading alternative architectures.
• Research and develop new applications of machine learning to improve our systems, such as self-supervised generative models utilizing our extensive unlabeled data.
• Build training and evaluation pipelines using field data, with particular attention to rare targets, imbalanced classes, and biased datasets.
• Define evaluation metrics and assess how models perform across target classes and field conditions, with particular attention to rare, high-impact events.
• Analyze model failures, data gaps, and changes in operating conditions to improve robustness.
• Partner with cross-functional teams to integrate models into production products.