Edge Engineer

1 hour ago

Vancouver, British Columbia, Canada Swim Recruiting Full-time

Join a small, high-impact engineering team building AI-powered computer vision infrastructure that's already deployed and running in live retail environments. This is a great opportunity to own meaningful, technically challenging work on the cutting edge of edge engineering, computer vision, and real-time business intelligence.

  • Take end-to-end ownership of our client’s edge stack, supporting the full on-device pipeline powering computer vision AI systems on high traffic live camera feeds.
  • You'd own a real system that is already in customers' stores, with clear and unsolved technical problems in front of you.
  • You'll be part of high impact team as our client enters an exciting period of growth. Your contributions matter immediately, and you will shape their technical direction.

What & Why:

A fast-growing retail AI company is looking for an Edge Engineer to own the on-device computer vision pipeline powering its in-store analytics platform. This hire is being made to deepen engineering capacity ahead of a significant multi-site expansion.

This position is hybrid 2 days on-site in Vancouver.

Who:

Our client is a small but proven technology company operating in the retail intelligence space, building edge AI infrastructure that runs entirely in-store on constrained hardware.

The team has a startup culture at its core — people here wear multiple hats, take ownership seriously, and move fast without sacrificing rigor.

You:

You will bring the following education, skills and experience to the role:

  • 2+ years of hands-on experience deploying computer vision or inference workloads to production edge hardware (NVIDIA Jetson or comparable embedded/accelerator platforms strongly preferred)
  • Solid Python skills, including comfort with performance-sensitive, production-quality code
  • Practical experience with a production inference runtime such as TensorRT, ONNX Runtime, OpenVINO, or TFLite
  • Working knowledge of multi-object tracking — association algorithms, IoU matching, Kalman filtering, and real-world failure modes like occlusion and crowding
  • Demonstrated rigor around evaluation: you've owned a benchmark harness, managed labeled datasets, or built an accuracy regression process, and you understand that measurement quality matters as much as model quality
  • Startup mindset — comfortable owning problems, working with limited resources, and contributing beyond a narrow job description

Compensation:

$120,000 to $150,000.