Database Reference
In-Depth Information
Gutierrez:
How do you think about whether you're solving the right
problem?
Heineike:
Well, that's a tricky question. I think that you have to iterate to get
there. With Quid, we have a lot of users who are asking questions that we
think are valuable ones to have answered well, so the question is then, are we
really helping them answer those questions well?
In this case, it's really good to keep the users continually in mind. Its good to
understand them well enough that you can then regularly use the product the way
they would, and see where it falls short, and make sure you're working towards
fixing that. For us that has also meant, does the product satisfy our own curiosity?
Can I quickly learn about Bitcoin or the microbiome or Apple using it?
It's definitely the case that it's really easy to get into weeds with the stuff, as
there are always thousands of options of different algorithms you could try
and different tweaks you could do. You have to work hard to stay focused on
the big picture.
Some of it is also having hunches about what's going to have the most value. I think
you have to go there and make those decisions. I think we've got a really good
team here, so there's a lot of people I can bounce ideas off of when I'm working
on things, or who have very strong ideas about what the right direction should be.
So it's really about making the most of the people around you as well.
Gutierrez:
Whose work is currently inspiring you?
Heineike:
My epic Twitter feed of the people that I think are interesting is
always inspiring to me. It's interesting during big sporting or political events to
see all the visualizations people construct to explore them. During the World
Cup, for example, Google's had their page. Twitter's definitely had a few pages
on it. All of those are fun to dig around and look at, for sure. Pete Warden and
his deep-learning SDK are also currently inspiring me. I think he's very good
at diving into some of the really hard problems and then doing something
really interesting and unexpected with it. So that's often an inspiring kind of
challenge—the challenge of thinking about what more could we be doing or
how we could be using what we're already doing a bit more creatively.
Gutierrez:
What does it take to do great data science work?
Heineike:
Data science is already kind of a broad church. There are a lot of
aspects that could call themselves “data science” or could fall under that label.
For me, the first thing you have to do is piece together this idea that “Here's
a really interesting problem, and here's data that could talk to that, and here's
the methodology that would take that data and do the right thing to it to
actually reach the output.” To me, the key is figuring out how you get those
three things—the right problem, the right data, and the right methodology—
to meld. So that's the first stage—just being able to envision how they come
together.
