We are a startup, and we will be honest about it
Lahza is an early-stage startup. We are small, we are funded by our first paying customer, and our budget is limited. We cannot match big-company salaries, and we will not pretend to.
What we can offer is what most jobs cannot: your models run on real cameras in a real restaurant within weeks, you own them, and you grow with the company. If you want a safe, well-defined job, this is not it. If you want to build something that matters and own it, read on.
What Lahza does
Lahza turns a restaurant’s existing cameras into a team of digital supervisors. A small box on site runs the vision models. Our cloud platform turns what they see into one live view of every area, WhatsApp alerts for each manager and a nightly summary for the owner, linked to the point of sale. We are live today with a restaurant group in Jordan.
Privacy comes first: no face recognition, and recordings stay on site.
What you will own
The edge boxes at client sites: NVIDIA Jetson setup, video pipelines (DeepStream or GStreamer), camera streams (RTSP) and time sync
Bringing a new restaurant live: camera angles, box sizing and stable streams, in days rather than weeks
The models: training and fine-tuning detection and tracking, exported to ONNX or TensorRT so they run fast on the box
Accuracy: checking every agent against real events on a regular schedule, then tuning until managers trust the alerts
Working with our platform engineer on how events reach the cloud, and guiding an intern on model evaluation
Who we are looking for
2+ years of hands-on computer vision: detection, tracking and video, beyond notebooks
Python, OpenCV and PyTorch, and exporting models to ONNX or TensorRT
Some hands-on edge or embedded work (Jetson, Raspberry Pi or similar), or a real drive to learn it fast
At home in Linux and Docker
An entrepreneurial streak: you have built something on your own initiative, or owned work far beyond your ticket
You care about impact: a model is done when a restaurant manager acts on it, not when the score looks good
Happy on site: cameras, cables and busy kitchens are part of the job
Clear communication in Arabic and English
Nice to have: DeepStream, Triton, multi-camera tracking, MQTT, or CCTV experience.
What we offer
Share options for the right person, after a trial period
Real ownership: the models and the edge boxes are yours
Direct work with the founder, and room to grow into the lead for vision and edge as the team grows
Experience you will not get elsewhere: your model’s output reaches a manager’s phone in a live business
Mostly remote, with regular days on site at client restaurants in Irbid
How we hire
Fast, and fair to your time:
A 30-minute call with the founder
A short technical task (3 to 5 days)
A decision within 48 hours of each step
How to apply
Apply with your CV and answer one question in a few lines: tell us about a vision model you got working outside a notebook (on a camera, a device or in production). What broke, and how did you fix it?
Job details are sourced from the employer's original posting.
Open job postingAbout the company
A staffing and consulting firm specializing in placing technology professionals in agile and software development roles.