Martin Andraud: Research work is much like doing sports

What kind of research do you do, and why?

My research is at the interface between integrated circuit (IC, also called microchip) design and artificial intelligence (AI). My main objective for the upcoming years is to develop energy-efficient AI circuits that can be embedded in the next generation of portable devices, such as smartphones, smartwatches, wristbands and medical devices.

Users want their devices to be specifically adapted to their needs and have different functionalities, and at the same time have long battery life. However, embedding more functionalities, such as AI, costs computing power and energy. Portable devices are inherently limited in both since they are small in size and typically run on batteries. My research aims to tackle this problem by finding more efficient AI algorithms and IC implementations.

How did you become a researcher?

I wasn’t particularly interested in research before my master’s studies. However, during an internship in a research lab in 2011 in Austria, I really got into finding new methods and helping the team to develop novel ideas. The more I discovered about research and IC design, the more I wanted to learn. Therefore, I decided to go for a PhD and it became clear to me that I wanted to work in academia. The freedom and the possibility to collaborate with people all over the world are some of the advantages of being a researcher. My interest in doing research has grown throughout my career!

What are the highlights of your career?

I did my PhD in France, in Grenoble, where I worked on applying machine-learning techniques to adapt analog and radio-frequency circuits. This gave me a taste of what it is like to combine AI and IC design, and I developed this further as a postdoc in the EU Horizon 2020 Future and Emerging Technologies (FET Open) project, Phoenix.

FET Open is meant for projects working on radically new future technologies and this type of project intends to bring innovative (and sometimes a bit crazy) ideas into research. We developed ultra-low-power intelligent sensor interfaces and data processing circuits, which can run on less than 1 uW of power while being adaptive through software and hardware optimization. The results have been published in reputable IC design journals, which is very rewarding. I genuinely enjoyed being part of the multidisciplinary project and it has given me many ideas for the future!

What is the most important quality for a researcher?

Doing research takes time, effort, and perseverance, but it offers a lot of freedom and possibilities. It is a lot like doing sports: you need to train and be committed to what you are doing, but you also need to have the passion to go training every day. Some days will be better than others—that’s part of it. In research, your big games are conferences, presentations, and journal papers.

It is really a mix between being rigorous and efficient in your work and being curious about trying new things that makes you a good researcher!

What do you expect from the future?

I hope that in the future we will be able to develop AI that really helps people and brings a benefit to society. I believe that, among other things, AI can help us to develop devices that are more energy efficient and serve our everyday needs better.

Martin Andraud started as the Assistant Professor at the Department of Electronics and Nanoengineering as of September 2019.