Convolutional Neural Network Architect

Are you a leader in the field of artificial intelligence algorithms, looking for an opportunity to change the entire field of AI and make a lasting impact on the world? Neurophos is looking for an individual who can and wants to be among the first to develop AI applications and algorithms for the first commercial optical computer. Neurophos exists to deliver an optical computer that will not only bring to AI a level of computational power that is not achievable with digital electronics, but also change how we think about computational resources, including how they work and scale.

Neurophos is hiring experts in neural network architectures to help us make this dream a reality. We are a fast-paced working environment with engineers and physicists with a diverse set of areas of expertise, and are always looking to add another team player who can deliver a valuable perspective that will help guide our product development and algorithmic capabilities into delivering the best that optical computing can deliver for AI.

Currently all work at Neurophos (outside of lab work) is done remotely during the COVID-19 pandemic. Plans are in place to relocate the team to Austin, TX once it is deemed safe enough for our employees.

Responsibilities include:

• Identify important operations and architectures for upcoming markets and applications, using CNNs.
• Contribute as needed in a dynamic startup company environment.
• Identify compute requirements for these operations and architectures.
• Develop a scheduler to break down a CNN architecture into low level hardware operations, given a hardware architecture.
• Perform network architecture search on neural networks for Neurophos’ optical computing platform.

Requirements:

• Proven track record of delivering, and drive to work in a variety of roles in an early stage startup.
• Bonus if 1+ years working with compilers and embedded systems.
• PhD in computer science, 3 years industry experience working in deep learning.
• 2+ years experience working specifically with CNN’s.