Internship: Model Predictive Control (MPC) running on a novel optical computer

Job Description

At Microsoft Research Cambridge, our scientists and engineers undertake world-leading, ambitious projects to transform the world through deep research. Our lab has contributed to many of Microsoft’s most successful products, services, and partnerships.

The Future AI Infrastructure team is a highly cross-disciplinary group of researchers and engineers with diverse expertise across computer science, optical sciences, image processing and engineering. We partner closely with product teams across Microsoft to tackle the big and open challenges of next-generation cloud infrastructure. We create new technologies and develop partnerships with leading academic groups across the world.

During this internship you will work with the researchers in Project Analogue Optical Computer (AOC). Additionally, as an intern you will be part of the larger internship program at Microsoft Research, Cambridge

Responsibilities

You will be responsible for creating a Multi-Joint dynamics with Contact (MuJoCo) Model Predictive Control (MPC), MJPC, controller that leverages the AOC optimization service for online computations, and applying this controller to control tasks implemented in MuJoCo. This includes:

(a) integrate the AOC optimization service into MJPC to create a standalone MPC controller;

(b) evaluate the MPC controller on selected optimization problems with respect to speed and solution quality.

Qualifications

Required/Minimum Qualifications:

• Studying towards a PhD in Physics, Mathematics, Computer Science, or related areas.

A successful candidate would have the following desirable experiences and capabilities:

• At least 1-3 years’ experience in designing and evaluating optimization algorithms for hard-optimization problems.

• Strong programming skills in a modern programming language. In particular, we are using the Julia programming language; experience in any similar language (e.g., Python, Java, C#/F#) and willingness to quickly learn Julia should suffice.

• Understanding of mathematical programming.

• Experience in mapping optimization problems to physical solvers.

Other Requirements:

Besides the core competencies above, a successful candidate would have many of the following desirable experiences and capabilities:

• Good organizational skills, capable of managing multiple simultaneous tasks effectively.

• Experience applying Computational Thinking techniques to solve problems in experimental science and/or obtaining reproducible results through algorithms and automation.

• Good presentation and communication skills.

• Prior knowledge in control theory and MPC.

Additional Qualifications:

• Experience in developing and maintaining software artifacts that are distributed as packages to a wider community.

Microsoft is an equal opportunity employer. Consistent with applicable law, all qualified applicants will receive consideration for employment without regard to age, ancestry, citizenship, color, family or medical care leave, gender identity or expression, genetic information, immigration status, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran or military status, race, ethnicity, religion, sex (including pregnancy), sexual orientation, or any other characteristic protected by applicable local laws, regulations and ordinances. If you need assistance and/or a reasonable accommodation due to a disability during the application process, read more about requesting accommodations.