Interesting job advertisement from Oculus (Meta / Facebook).
They seem to quite interested in neuromorphic sensing, with mentions of neuromorphic computing for AR and VR applications.
What is perhaps the most intriguing part is the term "neuromorphic time based" machine learning. which I suspect refers to time-series analysis in machine learning, and which is used alongside the terms "spiking neural networks" and "event-based sensing". Time-series analysis is commonly done with LSTM's and transformer networks, and as far as we know the only neuromorphic chip with these capabilities will be Akida 2.0 (which hasn't been announced as completed but the simulator is likely ready and possibly in the hands of customers already, as others here have pointed out).
If the above is true, it leads to the question: is Meta an EAP customer?
Keep in mind that this is a research role so if Brainchip is involved it will likely be at least a year or two till potential products hit the market (especially since Akida 2.0 isn't even in silicon yet).
Computer Vision Researcher - Reality Labs Research
Oculus
Full Job Description
Reality Labs Research (RL Research) brings together a world-class R&D team of researchers, developers, and engineers with the shared goal of developing AR and VR across the spectrum. The Surreal group at RL Research is seeking exceptional research scientists to solve the next generation of research challenges on the path to building
future machine perception enabled technologies. As a Computer Vision Research Scientist, your role is to collaborate with other researchers and engineers of the team in building novel approaches for
AR/VR based on
neuromorphic event based sensing technologies. This involves contributing to all steps of the development, including but not limited to,
developing novel event based algorithms, determining the
architecture and microarchitecture required to operate efficiently the approach,
verification, prototyping, and power/performance estimation.
Computer Vision Researcher - Reality Labs Research Responsibilities:
- Research, develop and prototype advanced novel event based vision algorithms and time based machine learning techniques
- Work with FPGA engineers to perform early prototyping of the developed methods
- Collaborate with team members throughout the lifetime of a project, from prototyping to deployed products
- Support integration of blocks with existing frameworks
Minimum Qualifications:
- PhD degree in Electrical Engineering or Computer Engineering
- 10+ years of research experience as a research scientist in the field of event based computation and neuromorphic engineering
- Experience in developing event based vision algorithms
- Experience in time based machine learning and spiking neural networks
- Experience working on multidisciplinary projects
- Interpersonal experience: cross-group and cross-culture collaboration
- Must obtain work authorization in the country of employment at the time of hire and maintain ongoing work authorization during employment
Preferred Qualifications:
- Experience with event based cameras
- Experience with neuromorphic time based development environments
- Tracking, visual odometry using event cameras experience
- MATLAB, Python, C/C++ or similar coding experience
