Deep Learning Software Developer
Algolux
Montréal, QC
As an Algolux Deep Learning Software Developer, you'll work on Computer Vision projects porting our core algorithms to embedded platforms. The ideal candidate can implement Convolutional Neural Network (CNN) architectures and scientific algorithms on selected hardware (e.g. GPU, multi-core ARM CPU, DSP) based on performance, power, and size constraints.

Key responsibilities

Implementation/Porting of TensorFlow algorithms to embedded automotive platforms. Practical experience with Nvidia toolchain for Jetson TX2 and Drive PX2 is desired.
Serve in a leadership capacity for carrying out embedded software development in C, C++, DSP Assembly or other selected languages.
Design and/or review architected modules and software systems supporting new technology or improving capability/performance of existing functionality.
Write application software on automotive DNN platforms, such as Nvidia, Movidius, Qualcomm and Renesas.
Research fundamental problems and implement algorithmic solutions that are appropriate.
Estimate scope, duration, effort and risk on project proposals.

Requirements:
TensorFlow, TensorRT and/or other embedded vision processing tools.
Optimization experience on embedded GPU and DSP platforms: Nvidia, ARM.
Parallelized & vectorized computation.
Solid GPU programming experience using CUDA, cuDNN, cuBLAS.
Excellent C, C++ programming skills.
Experience in signal processing algorithms and implementation.
Desirable

Experience in GPU programming using OpenCL.
Experience in Python programming.