AI Vision Research Engineer - Pico - San Jose

ByteDance ByteDance · Big Tech · San Jose, CA · R&D

Research Engineer focused on developing a power-efficient, low-latency, intelligent machine perception vision system for AR/VR products. This involves creating training and data pipelines, state-of-the-art ML algorithms for scene and user behavior understanding, mapping ML algorithms to hardware accelerators, optimizing performance, and collaborating on workload partitioning and hardware/software requirements. The role also includes quantitative evaluations using system modeling and FPGA prototyping.

What you'd actually do

  1. Innovate and develop a power-efficient, low-latency, highly-intelligent machine perception vision system in a VR/AR product.
  2. Develop necessary training, data collection/generation pipeline along with state of art ML algorithm/framework for understanding of the scene and user behaviors.
  3. Map the ML algorithm to the HW accelerator, identify performance bottleneck, optimize performance and efficiency, propose more efficient HW/SW co-design for AI accelerator.
  4. Collaborate with the x-functions team and translate system requirements into a proper cloud-edge workload partition and edge ML accelerator HW/SW requirement.
  5. Perform quantitative evaluations through the use of system modeling, FPGA prototyping for validating ideas and architecture options.

Skills

Required

  • C/C++
  • MATLAB
  • Python
  • ML algorithm research
  • AI accelerator design
  • System modeling
  • FPGA prototyping

Nice to have

  • Familiar with appropriate ML frameworks
  • Good communication skills
  • Open minded
  • Enthusiastic about new technologies

What the JD emphasized

  • PhD in the fields of Electrical Engineering, Computer Science, Physics and related technical discipline.
  • 3 + years of experience in research experience in the fields of ML algorithm or AI accelerator design
  • 3+ years for experience with programming tools and languages such as C/C++, MATLAB, Python.

Other signals

  • Develop ML algorithm for understanding of the scene and user behaviors
  • Map ML algorithm to HW accelerator
  • Optimize performance and efficiency for AI accelerator
  • Collaborate with x-functions team and translate system requirements into cloud-edge workload partition and edge ML accelerator HW/SW requirement
  • Perform quantitative evaluations through system modeling and FPGA prototyping