When it comes to creating a great product (say, an app that plays your favorite audio on demand), development requires patience, time, asking a lot of questions—and learning from failure. That’s where Tech Research at Spotify comes in. The team’s work allows us to develop core technologies and practices that ultimately enhance the user experience.
Mounia Lalmas-Roelleke is our Head of Tech Research for Personalization at Spotify. She has deep expertise leading interdisciplinary teams focused on user engagement in areas such as native advertising, digital media, social media, search, and now audio. We talked to Mounia about her team and its mission to incorporate and extend Spotify’s state-of-the-art technologies and products.
First off, why does Spotify need a Tech Research team in Personalization? What does it focus on?
I would take a step back and look at Spotify’s mission to unlock the potential of human creativity—by giving a million creative artists the opportunity to live off their art and billions of fans the opportunity to enjoy and be inspired by it. We believe this vision can be achieved by operating with state-of-the-art science and pushing beyond.
Within Personalization, we aim to match users with the content they want to hear at the right time, in the right place. This means getting the context right, in a way that is delightful. It’s about providing an experience that is special and fulfills user needs in the moment. But reading third-party books and papers is often not enough to solve everything we are tasked with, especially in the short timeline we are typically given. So we conduct our own research.
To do this, we focus on several areas—human-computer interaction, language technologies, machine learning, algorithmic bias, evaluation, search and recommendation, and user modeling, for example—that make personalization and discovery work for both our artists and listeners.
What is the Personalization Tech Research team make-up?
We are currently a team of around 20 research scientists, with more joining our ranks in the weeks and months ahead. Our team spans a diverse set of backgrounds, practices, and disciplines—we have and hire research scientists across areas ranging from computer science to social science.
Our team is based across the Spotify R&D hubs in New York, Boston, London, and Stockholm.
How does this team work with others at the company? How do you prioritize what needs to be accomplished?
Our working model is known as “embedding,” where our research scientists spend approximately 50% of their time in product teams. This is so they can learn infrastructure, teach state-of-the-art processes, and contribute to a product directly and in the moment. There is an immediacy, both in terms of time and impact.
In the other 50% of their time, research scientists investigate independent projects that serve as input for future product directions. This includes assessing the feasibility of new experiences, algorithms and methodologies, as well as deep dives into topics of interest that do not yet fit within existing work streams. These investigations can then lead to an embed to deploy that experimental work. We also use this time to contribute to the larger research community through publications and external presentations.
How does your function’s work help Spotify beyond R&D?
We believe that a strong internal research community and culture at Spotify will enable us to learn faster and have a bigger impact in our tech leadership. Our weekly Tech Research meeting, which is open to all employees, aims to strengthen Spotify’s internal research community and culture through sharing and discussing recent research from both inside and outside of Spotify.
What’s the key to Tech Research’s success?
It is important that as a team in Personalization, we adapt to the very rapid changes in the tech world, and are constantly learning and innovating. As long as we can do that, Tech Research will flourish.
For me personally, research means getting up every morning and being excited about the work I am doing, from building a hypothesis to seeing how it matures, to full deployment—for myself and anybody in my team and the teams who work with us.
Learn more about some of the Spotifiers working out of our London R&D Hub and follow us on @SpotifyJobs.