“We know music.” Q&A with Paul Lamere, Director of Developer Platform at the Echo NestDecember 02, 2011
by Alex Patriquin
Paul Lamere is a software developer with more than 30 years experience. He previously worked at Sun Microsystems and left to join the Echo Nest, a Somerville-based startup at the edge of innovation in music technology. Last month, Paul organized Music Hack Day Boston 2011 in Cambridge, which brought together a hundred or so developers and music geeks for an intense weekend of music hacks and a bumping party with none other than DJ Ali Shaheed from A Tribe Called Quest. Over the weekend, Paul and his daughter created the epic Bohemian Rhapsichord.
What does the Echo Nest do?
Our tag line is, “We know music.” We do everything we can to understand music. We do a deep crawl of the web, crawling every website where people write about music. So, every blog post, every music review, every biography, anything people write about music. We crawl, we pick on those words; we do natural English processing of all those words, and get a good understanding of what people are saying about music, which artists are hot now, which are artists are getting a lot of buzz. We can use this data to do cultural-based recommendations, and give people lots of context around musical artists.
The other thing we do is we listen to music. We run signal processing on this audio. We apply machine learning to all the features that you extract from the audio, and this gives a really deep understanding of what the music sounds like. So, for 15 million songs, we know where the drum beat is, we have a detailed view of the tempo and harmonic content. We know how danceable a song is, how energetic, how loud it is.
We take all this data, this cultural data, this audio data, and we combine them into a platform. Then we offer developer APIs so that people can call and build music apps on the platform. We also have some innovative play listing APIs so developers can build next-generation music exploration and discovery tools.
What do you at the Echo Nest?
As Director of Developer Platform, I define that strategy for the APIs. I get a feel for where the industry’s going, make sure the Echo Nest is there ready to serve the need. Today, we see people using music services like Spotify or Rdio that have essentially 10 million tracks to choose from. So, helping people wade through that is a big part of what we’ve been doing in the last year.
Longer term, I spend time trying to map out where we’re heading. We’re really well-integrated with a number of research communities; music informational retrieval community, there’s a recommendation systems community. So, we have an idea of what’s coming out of the world of research. We think a lot about how to take what’s coming out of the world of research and productize it, make it something that will help people who are tracking or listening to music.
How did you hear about the Echo Next? What prompted you to take a position with them?
So, Echo Nest was founded by two MIT Ph.D., Brian Whitman and Tristan Jehan, and both Brian and Tristan are pretty well-known in the music information retrieval research community. They had written some very fundamental and often cited papers. I knew of them by reputation. When they started the Echo Nest, naturally my curiosity was piqued. I followed them during their stealth startup mode as much as I could. Then once they started to hire people, I showed interest in joining the team.
What is your professional background?
My background is being a programmer. I’ve been writing software for years, and years, and years. I was pretty good at it, good enough so that I could work at a place like at Sun Microsystems; I was working in the research lab there. About five years before I joined the Echo Nest, that’s when I started a project in the research lab called Search Inside the Music which was, in some ways, very similar to what The Echo Nest has become. It was a project focused on helping understand music recommendation and discovery, where it was going and how we could help people deal with ever-growing music collections out there.
How does working at the Echo Nest compare with working at the labs at Sun Microsystems?
In some ways, they are similar. The type of work, the technologies were similar. However, there’s a very different pace. At a research lab, tasks are often very self-driven. I had the latitude to decide what I wanted to work on and spend time without a whole lot of schedule pressure. The downside of that is there was not a strong direct link from the work I was doing and product. And that was probably the main dissatisfaction I had working [at Sun]. The work I was doing was not critical to the success of the company.
So I felt that I wanted to have an impact in world of music technology, I needed to move to a company that had more skin in the game. The Echo Nest was right there for doing all of that.
Were there any challenges in making the transition?
The biggest difference is the pace and urgency here at The Echo Nest. We have. There’s such a large interest from so many companies in what we’re doing that we spend. It seems like every day we have a new exciting project to work on. Sometimes it can be overwhelming, the amount of things that are in the queue, but at the same time it’s really heartening to know that people are really interested in what we’re doing and want to use it.
Would you recommend your former colleagues at Sun consider startups?
Oh absolutely, yeah. There’s nothing like doing something on Wednesday and seeing Thursday that it’s the top of Hacker News, the top of Reddit, and the top story on TechCrunch. The amount of impact my work has on the company, there’s no comparison. At Sun, I met the CEO once and it was kind of a big day for me. Here the CEO and I sit down every day. So there’s quite a big difference in the amount of distance between me and the top of the company. If you want to have an impact, being in a startup is the place to be.
Great. One final question. Are you hiring right now?
Yes, we are. If you go to http://the.echonest.com/company/jobs/, we’ve got a bunch of openings. We’re looking for very smart engineers, especially those who understand scaling. We also have openings for audio processing engineers.