Database Reference
In-Depth Information
Gutierrez: How do you work with your team members?
Lenaghan: Once the teams arrive we have a morning standup with the
engineering and the data science teams around 10…10:15. Everybody talks
about what they are doing. I then sync up with a few team members through-
out the day and lead or attend a variety of meetings. The meetings are usually
product meetings or troubleshooting meetings. The troubleshooting meetings
come up if we are having an issue with a particular campaign we are serving.
These one-on-one and small group meetings go on throughout the day.
In between meetings, I am usually in discussions with account managers. These
discussions center around which campaigns are feasible and which are not.
What sort of targeting could be used? And is this type of campaign and tar-
geting even possible? I will get questions whether a particular audience—even
though it sounds very interesting—is going to have the necessary scale to
make an impact. I find myself having to make a lot of higher-level strategic
decisions. I try to keep most members of my team as separate from those as
possible so that they can really focus on the clients, projects, and data. Then
around 6 o'clock my day winds down and I go home.
Gutierrez: When thinking through and making these higher-level decisions,
how do you think about whether you and your team are solving the right
problems?
Lenaghan: I always try to look at the problem from the end. So I think about
what is the final output and functionality that we want after all the days or
weeks of work have been put into solving the problem. Is the final output a
particular audience? Or do we want a classifier to perform much better? Or
perhaps we already have a process, but the machine learning component of it
is not performing as well as we would like and we want to improve it. I always
start from the end.
What I have found—not only from working in industry, but academics as
well—is that when you start from the beginning and everything is blue sky,
there are hundreds of ideas to chase as well as thousands of ideas to try and,
since everything is possible, nothing ever gets done. It can and has happened
that things eventually get done, but running a company by serendipity is beg-
ging to fail. So I always focus on looking towards the end result.
Of course, many times throughout the course of solving the problem, you
end up at a different place. Sometimes it is better; other times you just have
to scrap the project. Keeping your eyes on the final deliverable is essential to
solving the right problems.
Gutierrez: There is an idea in engineering that you build one to throw away.
Do you find there is a sense of that in the data modeling work that you do?
Lenaghan: I definitely think so.
 
Search WWH ::




Custom Search