Bookmark and Share

Jobs Available/Wanted

Embedded Video Processing and Machine Learning Engineer
No replies
Phelps2020
Phelps2020's picture
Offline
Last seen: 7 weeks 6 days ago
Level 1: Prestidigitator
Joined: 2019-09-01
Points: 1

Company Background:

Phelps2020 is a fast-growing company that develops and commercializes cutting edge embedded computer vision solutions for use in military and commercial applications. We are looking for a curious self-starter who is passionate about tackling tough technical problems. Candidates should be excited about product development and eager to take on roles and tasks that challenge their current technical knowledge base.

Position Background:

The candidate that takes on this role will be responsible for embedding video processing and convolutional neural networks (CNNs) onto Xilinx Zynq® UltraScale+™ MPSoCs. The candidate should be very comfortable ingesting and manipulating high-throughput live video on the above Xilinx systems. The candidate should also be comfortable optimizing and embedding CNN inference models onto Zynq® UltraScale+™ MPSoC using Vivado, SDSoC or DNNDK. He or she should also be comfortable working with a team, reaching out to the community for guidance, managing/advising 1-3 technical resources and sharing their findings with other engineers working on complimentary efforts.

Special Consideration:

Due to government contracts held by Phelps2020, all employees must be US citizens or have permanent residency.

Responsibilities Include:

  • You will be responsible for embedding software to perform video ingestion across multiple interfaces: ethernet, SDI, HDMI, etc.
  • You will be responsible for embedding software to perform various image processing routines on live video up to 4K 60FPS
  • You will be responsible for optimizing and transferring CNN inference models to Zynq® UltraScale+™ MPSoCto perform object detection and classification
  • You will be responsible for system architecture, micro architecture, RTL coding, IP selection, IP evaluation, IP integration, building, verification and hardware bring up on Zynq® UltraScale+™ MPSoC.
  • You will get to work closely with our image processing and machine learning team to understand and define the functional and performance goals of the design. Work independently & handle deliverables to align with the project goals plus support cross-functional engineering efforts.

Key Qualifications:

  • BS in Electrical Engineering, Computer Engineering (EE-CE) or Computer Science from an accredited institution with 3-5 years of related experience, or MS degree in EE-CE with 1-3 years of related experience, or PhD in EE-CE in a related sub-field
  • 2+ years of experience with the latest Xilinx Zynq® UltraScale+™ MPSoC technologies as well as 7 series
  • Strong Verilog or VHDL RTL design skills
  • Strong SDSOC and Vivado HLS experience
  • Experience with DNNDK framework
  • Experience with video industry-standard interfaces/protocols (PCI Express, Ethernet, DP, LVDS, HDMI, MIPI, Optical etc.) are pluses
  • Experience with high speed design, including DDR3, SERDES, Gigabit Transceivers, PCle, Ethernet, etc.
  • Experience with video processing (camera interfaces, ISP, color, scaling, compression, etc.)
  • Understanding of Tensorflow/Caffee convolutional neural networks and the process by which they are embedded onto FPGAs
  • Experience with MATLAB Simulink, C++, C, Python and computer vision libraries such as OpenCV
  • Knowledge of embedded Linux/PetaLinux

Location:

  • Knoxville, TN is preferred, but remote candidates will be considered if they are strong communicators and capable with video conferencing.

Job Type: Full-time