Our client company is seeking a deep learning software architect for building neural network models suited with their ONN accelerator.
? Develop software API that controls the ONN, along with its dataflow, using standard deep learning packages such as Tensorflow, MXNET, or Theano.
? Work with hardware engineers on engineering efficient high-bandwidth dataflow
? PhD (MS, minimum) computer science, physics, or related field
? Highly proficient in machine learning programming languages and packages, e.g. Python, C++, Tensorflow, Theano, MXNET. Preferably with 3+ years of experience in low-level coding of machine learning languages.
? Experience in maintaining a large codebase with good software engineering habits.
? Experience with real-time constraints, and hardware interfaces (e.g. Ethernet, UART, SPI, I2C)
? Strong background in data structures and algorithms
? Understanding of the different machine learning accelerator architectures, e.g. GPU, ASIC, FPGA, ONN.