FR - Machine Learning Engineer

Position: Machine Learning Engineer

Location: France - Nice

Reporting: VP ML/AI

Role: Full-time

The Position

MARSS is investing in the Machine Learning field in order to improve our NiDAR software in the following areas:

• Video Classification, Detection, Tracking


• Pattern recognition / Anomaly detection


• Radar tracking


• Radio Frequency analysis

Our current focus is strongly in the video area, but we aim at growing quickly our experiments in the other fields.

Our team is growing, and we expect the Machine Learning activity to grow the most.

Reporting to our VP ML/AI, we expect the new hire to:

• help collecting and categorizing data from public sources but also our customer base (when authorized)


• help with training data - we outsource annotation, but data still needs to be validated and sometimes annotation transformed for the appropriate network; 


• in particular, we build our own tools to help filter and validate the data and the training results, and we expect to keep improving


• test performance and compatibility of new hardware, not only in the Nvidia world with the edge Jetson platform, but also on programmable FPGA or pure server-grade CPU with Intel


• test, improve, and tune existing algorithms (neural networks, Bayes/Markov, decition trees, reinforcement networks, Gaussians...)


• convert existing solutions (and training data) to different frameworks

Requirements 

• We are technology agnostic but favor TensorFlow, Keras, and PyTorch; 
  


• Good Python knowledge is important both for the data analysis but also for understanding integration into our backend (which is python based).
  


• We work on Linux (although laptops are Mac, which is now a second class citizen to Nvidia...)
  


• We require either working experience or at least appropriate theoretical knowledge (or both).

Education

We prefer at least bachelors degree level in a science or engineering discipline. Masters or doctorate advantageous.

Integration and interfacing

Our core activity is maritime safety and security applications. We're bringing the latest web technologies to real-world hardware i.e. we don't just work in the cloud. We interface with very diverse equipment: radar, sonar, thermal imaging, software-defined radios and more. Our UI runs on touchscreen displays, tablets, phones, smart-glasses...

Challenges

If you are looking for an easy job, this isn't it. We operate in mission critical environments. Our systems can operate at sea and high security areas with limited connectivity to the data center. So, all your assumptions about cloud development techniques and operating environments will be constantly challenged. Timelines are strict, and creativity is often required to solve unexpected technical issues. It can be extremely diverse, interesting and rewarding if you are up to the challenge.

The Work Environment

• Development on Macs and Ubuntu: We're pragmatic so realize we also have to work with Windows occasionally.


• Prestigious location: Overlooking the Monaco Grand Prix track and port.


• Outdoor sport paradise: Nearby mountains and skiing, cycling opportunities, and obviously the sea
  


• Small team: Embracing modern development practices. New ideas are always encouraged and we'll try most things, but we always settle on what works for us.


• Diverse team (in a good way): Our local team in Monaco is very multicultural including experienced engineers balanced with the brightest young graduates.


• Passionate: We are passionate and enjoy what we do!