Cambridge Residency Programme : Machine Learning for Signal Processing

Key Skills

Analytical and Problem solving, Computer Vision (CV), Data Analysis, Deep Learning, Machine learning techniques, PyTorch, Signal processing, TensorFlow

Job Description

Machine learning is having a transformative effect on many research areas, fundamentally altering the way many problems are approached, and the scope of what is achievable.

Project Silica is developing a revolutionary storage technology by using femtosecond lasers to store data in glass. The project provides an unprecedented opportunity to completely re-think how storage systems are built, and build a storage technology that’s end-to-end optimized solely for the cloud, from the materials level up to the user interface.

Machine learning plays a critical role in the development and existence of the technology, enabling unprecedented information densities to be recorded and successfully read back from glass, facilitating end-to-end optimization of the write and read processes across hardware and software, and more.

The Project Silica team is highly multi-disciplinary, comprising computer scientists, physicists, optical scientists, chemists, electrical and mechanical engineers, and designers. Tackling the hard technical challenges of making glass storage into a real technology presents a myriad of unique opportunities for cross-disciplinary research, including applying machine learning to completely new problem domains.

Contract duration: 2 years.

Qualifications

Required/Minimum Qualifications: 

• Completed (or on-track to complete) a PhD in a relevant discipline.
• Knowledge of deep learning.
• Interest in applied research with real high-world impact.
• Strong software engineering skills and data analysis for rapid and accurate development.
• Creative and collaborative approach to problem solving.

Preferred/Additional Qualifications:

• Experience optimizing ML [particularly vision] models towards running in production on inference-optimized hardware.
• Publications in top tier conferences.
• Hands-on experience with current deep learning frameworks (e.g., PyTorch, Tensorflow, etc.).

Responsibilities

As a researcher on the project, you will be empowered to collaborate with various members of the team, designing, implementing, and applying machine learning solutions for signal processing to decode data read from glass, with the goal of making Silica into a successful storage technology deployed at scale in the cloud. Problem areas broadly include: computer vision for object detection and recognition, optimization of machine learning models towards running in production, ML-driven discovery and optimization of new physical processes, and more.