About Us:
Swan delivers the future of defense and industry through scalable autonomous products. Swan is backed by leading defense VCs including a16z's American Dynamism fund.
Role Overview:
We are seeking an expert Machine Learning Engineer with deep experience in computer vision, model optimization, and deployment on low-cost embedded systems. The ideal candidate will have a strong background in designing, training, and optimizing deep learning models for real-time applications. This role requires expertise in efficient neural network architectures, quantization, model compression, and hardware acceleration techniques to run ML models on resource-constrained devices.
Key Responsibilities:
Design, develop, and optimize computer vision models for real-time applications on embedded systems.
Implement model compression techniques such as quantization, pruning, and knowledge distillation to improve performance on low-power hardware.
Deploy machine learning models on embedded platforms, including ARM, NVIDIA Jetson, Qualcomm, or custom ASICs.
Write clean, efficient, and well-documented code in Python and C++, leveraging ML frameworks like TensorFlow, PyTorch, and ONNX.
Develop and fine-tune SLAM, object detection, tracking, and feature extraction models for high efficiency.
Collaborate with cross-functional teams to integrate ML models into production systems, optimizing for latency, accuracy, and power consumption.
Benchmark and profile ML models to identify and implement optimizations for inference on embedded hardware.
Research and apply cutting-edge ML techniques to improve real-time performance in resource-constrained environments.
Qualifications and Skills:
Master’s or Ph.D. in Computer Science, Electrical Engineering, Machine Learning, or a related field.
5+ years of experience in machine learning, deep learning, and computer vision.
Extensive experience in designing and deploying optimized deep learning models for real-world applications.
Proficiency in TensorFlow, PyTorch, ONNX, TensorRT, and other ML frameworks.
Strong experience with model quantization, pruning, knowledge distillation, and hardware acceleration techniques.
Solid programming skills in Python and C++, with a strong understanding of software optimization.
Familiarity with embedded platforms such as NVIDIA Jetson, Raspberry Pi, ARM Cortex, Qualcomm AI accelerators, or specialized AI chips.
Experience with hardware-aware model optimization to maximize inference speed and minimize memory footprint.
Strong problem-solving skills and ability to work independently on complex technical challenges.
Preferred Qualifications:
Experience with real-time SLAM, visual odometry, and multi-sensor fusion.
Knowledge of low-level hardware acceleration using CUDA, OpenCL, or specialized ML accelerators.
Familiarity with robotics frameworks such as ROS for integrating ML models into robotic systems.
Background in edge AI deployments and optimizing neural networks for mobile and IoT devices.
What We Offer:
Swan is a fast-paced, innovative, and collaborative startup environment. We offer a top-notch benefits package, including:
Top-tier health, dental, vision, short-/long-term disability, and life insurance, with full employee coverage and partial coverage for dependents
Flexible/reasonable vacation and sick leave
401(k) plans, FSA, HSA, and commuter benefits
Swan is an equal opportunity employer and values diversity. We do not discriminate on the basis of race, color, national origin, creed, religion, sex, gender, gender expression, sexual orientation, age, marital status, uniform service, veteran status, disability status, or any other protected characteristic per federal, state, or local law. Swan considers qualified applicants with criminal histories, consistent with applicable federal, state and local law. We are also committed to providing reasonable accommodations for qualified individuals with disabilities and disabled veterans in our job application procedures, and we work to create a welcoming and supportive environment for all applicants throughout the interview process.