“The eye sees only what the mind is prepared to comprehend.”
Robertson Davies
The 3 skills that matter most
- Computer vision in production. Detection, tracking and re-identification models running on real cameras in the real world, not in a notebook.
- Elixir, or another functional language. Haskell, Scala, Clojure, F# and OCaml all count.
- Python. For inference, model evaluation and training pipelines.
The opportunity
Most businesses have a camera on the wall, and almost none of them use what it sees. The footage sits on a hard drive until something goes wrong.
Our client is changing that. They’re a venture-backed applied AI startup that turns ordinary store cameras into live operational data for restaurant, retail and convenience brands. Their platform already runs in thousands of locations, and new enterprise contracts are about to multiply that footprint many times over.
They need a senior engineer who can make computer vision work reliably at that scale.
This is applied vision engineering, not research. You won’t be inventing new architectures. You’ll take models that work in the lab and make them hold up in 3,000 different stores, each with different lighting, camera angles, glare, crowds and drive-thru lanes.
What you’ll build
- People analytics: re-identification, tracking across cameras, telling staff from customers, and detecting groups.
- Vehicle analytics for drive-thru lanes: detection, counting, timing and queue behavior.
- Model monitoring: systems that measure accuracy at every new site, flag when it drifts below standard, and feed that back into retraining.
- Vision tooling and testing: harnesses for trying new model variants on edge devices and in the cloud before they ship.
- The Elixir backbone: from video ingestion and efficient inference through to storing results, triggering alerts and running customer-facing workflows.
- Install standards: camera and hardware requirements that installation partners follow, so large rollouts go smoothly the first time.
The tech
- Elixir and OTP for the core platform: concurrency, fault tolerance and real-time pipelines.
- Python for inference, model evaluation and training workflows.
- Edge devices plus cloud inference, so you’ll be balancing latency, bandwidth and cost.
- Video ingestion at scale across thousands of live camera feeds.
- An AI-first engineering culture. The team uses coding agents and modern AI tools every day and expects you to as well.
What you bring
- 4+ years as a software engineer, including real time at a venture-backed startup.
- At least one computer vision or ML system you took from concept to production yourself, with evaluations, training pipelines and real-world tuning.
- Hands-on work with detection (people or vehicles), tracking or re-identification. Camera, surveillance, smart-city, retail-vision or autonomous-checkout experience is the ideal background.
- Production Elixir, or solid experience in another functional language and a real appetite to go deep on Elixir.
- Strong Python for the ML side.
- Heavy, curious use of AI tools. Open-source work or side projects are a big plus.
- A builder’s attitude. People who don’t fit the usual mould are welcome, including self-taught and bootcamp engineers with standout projects.
This probably isn’t for you if
- your background is pure ML research or LLM-only, or you mainly train models from scratch
- your experience is government or defense
- you’re a manager or executive who’s no longer coding every day
- you’re a strong web or backend engineer who has never shipped computer vision or ML to production
Why this role
- The timing is right. A proven product heading into a big enterprise expansion, so your work will run in thousands of real locations.
- A market almost nobody else is working on. Over 70% of businesses have never used their on-site video. This is a chance to help build the category leader.
- Real ownership. A small, senior engineering team with no layers. You own vision from model to production.
- Remote with structure. Fully remote in the US, with at least 5 hours overlap with Mountain Time and one team on-site visit a quarter.
Package
- $185K to $205K base plus meaningful equity
- Top-tier health benefits and flexible PTO
- Remote anywhere in the United States
- You must be a US citizen or Green Card holder. Visa sponsorship isn’t available.
Interview process
An intro chat, a collaborative product-scoping session, then a technical deep dive with the team.
Ready to apply?
Send your resume along with:
- A computer vision or ML system you took to production, and how you kept it reliable on messy real-world data.
- Your experience with Elixir or another functional language, and what you shipped with it.
- Your work authorization and location.
- Your salary expectations.
Thousands of cameras are watching. Help them understand what they see.
