Senior Software Engineer, Video Analytics

About the company

Our client is a Series A computer vision company of around thirty-five people, headquartered in Utah, that brings AI video analytics to thousands of restaurant, convenience-store and retail locations. It has raised roughly $19M, and it has recently moved upmarket: new enterprise deals in the pipeline represent something like twenty times its current site count, with rollouts running to three thousand sites and beyond. The entire stack is Elixir, from video ingest and inference through to alerting, and the engineering team is deliberately small — a lead, a product manager, two senior engineers and one mid-level engineer.

The role

You would own applied computer vision work that live stores depend on: people analytics such as person re-identification, cross-camera tracking, employee tagging and group detection, plus drive-thru vehicle analytics. That means per-store model performance monitoring, the evaluation and training pipelines behind it, the vision tooling and test infrastructure around it, and both edge and cloud variants of each model. The hiring manager is direct about the gap he is hiring into: engineers who genuinely bridge engineering and applied vision at startup pace. This is applied vision rather than research, and the work is judged on whether it stays reliable on messy real-world footage.

What you'll do

• Build and ship applied computer vision features that thousands of live store locations depend on
• Own people analytics: person re-identification, cross-camera tracking, employee tagging and group detection
• Build drive-thru vehicle analytics, from detection through to the metrics the customer sees
• Monitor per-store model performance and refine the signal when real-world footage degrades it
• Build the evaluation and training pipelines, vision tooling and test infrastructure behind the models
• Ship both edge and cloud variants of each model
• Work across an Elixir stack that runs from video ingest and inference to alerting

Nice to have

• Open-source work or AI side projects, and being a heavy user of AI tooling
• Cloud video, streaming or surveillance experience — the hiring manager notes that most candidates do not have it
• A bootcamp or self-taught background, if it comes with standout projects

Tech stack

Elixir end to end — video ingest, inference, alerting and the services in between — so hands-on Elixir, or a strong foundation in another functional language, is a core requirement rather than a preference. Around it sits applied computer vision on edge devices and in the cloud: detection and tracking models, evaluation and training pipelines, and the monitoring that catches one store's model drifting while the rest hold.

Compensation & logistics

• $185k–$205k base, plus equity
• Fully remote from the United States, with five or more hours of overlap with Mountain Time
• One on-site visit per quarter to the company's office in Farmington, Utah
• The client cannot sponsor or transfer a visa, so you need to be legally authorized to work in the US and not need visa sponsorship or a visa transfer now or in the future
• Two hires on this role, and the client wants to hire within about a month
• Interview process: a 30-minute initial screen that is mostly behavioural, a 60-minute collaborative product-spec exercise with the lead engineer that tests product sense and spec writing rather than system design, a 45- to 60-minute team interview combining a technical deep dive with behavioural questions, an optional final conversation with the CTO, then references
• Full-time