Computer Vision Sensing Engineer

  • 149 W 36th St, New York, NY 10018, USA
  • Full-time

Company Description

Perch is the retail digital engagement platform that detects which products shoppers are touching at the shelf and respond with product-level media and applications on integrated touchscreens. Our in-store shopper marketing platform drives 30-80% sales lifts from leading brands and retailers including Johnson & Johnson, Purina, MAC, Macy’s, kate spade new york, Jo Malone and Invisalign, and have just begun closing seven figure deals. We are a pre- Series A startup looking to grow our technology team and leadership as we look to scale up in 2021.

Job Description

Our computer vision sensing engineer is tasked with leading the advancement of our retail behavior sensing libraries.  Our current capabilities are functional and delivering transformative but basic, compared to the breadth of our vision.  We are tasked with improving reliability, specificity, and detecting new behaviors through computer vision interpretation (where people are touching, what they are touching, sensing pickup or walking away with a product) and other sensors (bluetooth for presence), local libraries for demographic segmentation.  Our goal is to deploy these libraries distributed at scale on commoditized hardware that may not be continuously connected to the Internet.  This role is for someone who can get their hands dirty and develop the algorithms while balancing the disciplinary concerns of sensor types, hardware costs, reliability and network variation.   When successful, you have the opportunity to change how in-store digital marketing works in a $4.4 trillion industry (in-store retail).   

Qualifications

  • Expert in C++/C and Python programming
  • Demonstrated skills in writing a production ready highly efficient code that run on embedded systems
  • Deep understanding in traditional computer vision methods
  • Experience with common deep learning libraries such as TensorFlow, PyTorch
  • familiarity with docker and docker-compose
  • Worked in a startup environment. You’re comfortable with ambiguity and change.
  • Real time video processing experience including semantic segmentation, object detection and object tracking.

Additional Information

All your information will be kept confidential according to EEO guidelines.

Videos To Watch