A complete application in a minute
Tailored resume and cover letter, ready to send.
Sorontar is building a universal weapon architecture and seeks an engineer to develop and optimize real-time computer vision algorithms for embedded hardware, guiding weapon systems through their phases.
You will train and validate models during tests, with a focus on performance within strict power and latency constraints, and collaborate in both French and English.
Sorontar is a team of engineers and scientists working to rearm France and its allies. Backed by world-class defense investors and angels, we are building a universal weapon architecture where software, design and manufacturing come together to produce more, faster and at lower cost.
A product like this calls for a different kind of company. In most large organizations, layers of management stand between people and their work: ownership fades, knowledge travels slowly and energy goes into coordination. We chose the opposite. We keep Sorontar flat and give that energy back to the work, which is why we look for people who need no management once the direction is set.
If a lean team driven by its mission is where you do your best work, you will feel at home here.
At Sorontar, you will work within the GNC team on the vision systems that guide our weapon systems through their different phases. You will develop the algorithms that detect, recognize and track a target in real time, and that help the system locate itself. These algorithms run on embedded hardware with strict limits on computing power, energy and latency, so you will design them with these constraints in mind from the start.
In the beginning, your work will be largely focused on our tests, where you will train and validate your models.
You have developed computer vision algorithms for detection, recognition or tracking that work in real conditions. You are comfortable with both classical methods and deep learning, and you know how to make a model run in real time on embedded hardware. You write solid code in Python and C++, and you work in French and English.
Having designed and trained image recognition models from scratch will set you apart. More broadly, experience with models that combine very different sensor inputs is a strong plus, especially in embedded and mobile systems. Experience in defense counts too.
We give priority to candidates with at least a first professional experience, and for candidates early in their career, we look closely at degrees and training.
People who thrive at Sorontar also share four traits:
We are looking for brilliant, fearless and good people, and we deeply believe that people do not fit into boxes. Yet while a person is a continuum of qualities, a hiring decision is a yes or a no. To make the best decision possible, we assess every candidate on six criteria, grouped in pairs in the matrices below.
Make sure you see the graphs on our career page: here.
Brilliant. Talent without responsibility leaves potential unused, and responsibility without talent reaches its limits quickly. We look for people who combine the ability to solve hard problems with the drive to see them through.
Fearless. Courage without judgment leads to reckless decisions, and judgment without courage turns into commentary. We see people who see clearly and act on what we see.
Good. Integrity without warmth makes a team hard to work in, and warmth without integrity is charm that does not last. We look for people others can rely on and enjoy working with.
The best companies are the ones where you work alongside the best colleagues. Our first commitment is to surround you with people you learn from every day, which is why we hire with great care.