Cambridge Residency Programme – Computation in Deep Learning

Job Description

Overview

Come and work with us to shape the future of computation in Deep Learning.

The Helia team in Microsoft Research Cambridge (UK) invites researchers with a focus on efficient computation to join our two-year Residency Program.

Helia is a research project with the goal of transforming how computation works for Deep Learning. By combining a mathematical understanding of deep learning models with hardware and systems knowledge, we develop new methods that enable efficient training and deployment of AI systems, scaling from datacenter-powered AI models to on-device models.

Our team collaborates with hardware development teams, AI deployment teams and research teams with complementary expertise to develop the very best methods for efficient AI. We have active collaborations with teams working on future hardware, including that shipped as Windows Copilot. You can see some of our recent work at top venues here and here.

This is an exceptional opportunity to drive ambitious machine learning research while collaborating with domain experts to push the state of the art in efficient AI, with the potential for enormous real-world impact.

We welcome applicants with either a Deep Learning background or an equivalent technical background wishing to move into AI, for example mathematics, physics or engineering.

Responsibilities

• Develop and drive an ambitious research agenda in efficient AI.

• Work closely in a team of machine learning researchers and software engineers to develop cutting edge methods for efficiency in AI training and deployment.

• Collaborate with teams across Microsoft to co-develop methods that improve efficiency across the AI stack, from hardware development through systems and deployment.

Qualifications

Required:

• PhD in machine learning, statistics, or a related field, or equivalent experience.

• Strong software skills with a deep learning framework: PyTorch or equivalent.

• Experience of best-practice research software development, including version control and unit-testing.

• The ability to think creatively about deep learning and computation.

• An understanding of the computational implications of deep learning operations.

• A growth mindset, curiosity about learning new things and a desire to deeply understand AI systems from a first-principles view.
  

Preferred:

• Publication record in relevant conferences, such as NeurIPS, ICLR, ICML, etc.

• Expertise in a mathematical aspect of deep learning, for example ODEs, statistical learning theory, optimization, information theory.

• Understanding of quantization, customized floating-point representations, compression and coding theory.

• Experience with CUDA development or equivalent.

Microsoft is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to age, ancestry, color, family or medical care leave, gender identity or expression, genetic information, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran status, race, religion, sex (including pregnancy), sexual orientation, or any other characteristic protected by applicable laws, regulations and ordinances. If you need assistance and/or a reasonable accommodation due to a disability during the application or the recruiting process, please send a request via the Accommodation request form.

Benefits/perks listed below may vary depending on the nature of your employment with Microsoft and the country where you work.