Embedded Machine Learning (ML) Engineer

: About Us: Cognex is a global leader in the exciting and growing field of machine vision. Our employees, proudly called “Cognoids”, are passionate about solving the most difficult vision problems with our embedded cameras and software products featuring state-of-the-art 2D and 3D vision technology using deep learning. Our Work Hard, Play Hard, Move Fast culture recognizes achievement and dedication with unique rewards and celebrations. This position is hybrid in our Natick, MA corporate HQ - 3 days in the office and 2 remote. We are looking for creative, bright, motivated Cognoids who share our passion for excellence and want to make an impact at a dynamic, global company. We celebrate our employees for their innovation, perseverance, and hard work in a fun, rewarding, and quirky environment. If you enjoy the sense of accomplishment that comes from working together to create products that solve tough problems for organizations around the world, contact us to see how you can become part of our team! The Team: Deep Learning R&D - Advanced Vision Technology This position is for an Embedded DL Engineer in the AI R&D group in the Advanced Vision Technology organization. This team is responsible for researching, designing, implementing and deploying state-of-the-art deep learning algorithms for industrial machine vision applications, with a mission to innovate on behalf of customers and make this technology accessible to a broad range of users and platforms. We apply the latest technology in computer vision to solve real world problems in the manufacturing industry such as defect inspection, barcode decoding, robotic guidance, image generation, etc. We have a particular focus on developing AI solutions that can run on embedded systems, such as Cognex's smart cameras. Within the team, development is done in Python and C++. Engineers in this group typically have experience with deep learning, image analysis, machine vision, statistical learning, or signal processing. The Role: We are seeking an engineer passionate about embedded and power-efficient deep learning. You will be working within the team, but with a special focus on how to run the neural networks we develop efficiently and accurately on embedded devices, both on ARM CPUs as well as on neural network acceleration ASICs from various vendors. As you grow in the role, we hope that you grow to be a go-to expert on these topics. Essential Functions:

• Optimize neural networks to run efficiently on embedded devices
• Evaluate neural network hardware accelerators, including working with the various vendors' custom hardware stacks
• Research and evaluate methods to run neural networks more efficiently, such as quantization or model pruning
• Share your knowledge about efficient ML with the broader deep learning R&D team, to help us design networks and algorithms that are more efficient on low-power platforms

Knowledge, Skills, and Abilities:

• Experience in industry or academia experience focused on efficient deep learning
• Experience with at least one neural network inference framework such as TFLite, ONNXRuntime, or TVM, including some understanding of how they work internally
• Experience with topics such as neural network quantization, pruning, or compression
• Strong programming skills in C++
• Experience with SIMD programming or other optimization techniques is a plus
• Experience with neural network accelerator ASICs is a plus

Minimum education and work experience required:

• Bachelor's degree in CS, EE, or equivalent

Additional Equal Employment Opportunity Cognex is an equal opportunity employer. Cognex evaluates qualified applicants without regard to race, color, religion, gender, national origin, age, sexual orientation, gender identity or expression, protected veteran status, disability/handicap status or any other legally protected characteristic.