What is your role at GridCure?

I’m Head of Analytics.  I spend quite a bit of time researching the utility industry, working with clients, and developing predictive modeling solutions to help utilities make better use of the data that is being collected on the smart grid.

What drew you to Gridcure originally? And how has GridCure changed since?

I thought the role would be a good fit for my background.  I have a Ph.D. in applied mathematics from Northwestern.  The applied math department there is in the engineering school, so I had a fair amount of interaction with engineering departments while I was there.  Applying mathematical modeling to engineering problems comes naturally to me.

Because of the expansion of the smart grid, the amount of data available to utilities has increased by several orders of magnitude over the last decade.  The influx of new data means that there is quite a bit of low-lying fruit as far as the problems that need to be solved in the industry.  Utility analytics is a very exciting field at the moment.

GridCure has changed quite a bit since I started working here in October of 2014.  When I first started the company consisted of four of us in different locations joining a google hangout for meetings twice per week, and working mostly on one project.  Since then, we’ve participated in the 500 Startups program in San Francisco and set up offices in three locations.   We’ve been receiving increasing interest from utilities and are working on a very diverse range of projects.

Can you tell us a bit about the competitive predictive modeling you’ve done?

Kaggle is the most respected competitive predictive modeling site in the world.  I joined Kaggle about a year after the site was founded and started working on some of the predictive modeling projects there.  It’s an excellent way to get experience doing predictive modeling for large  real world datasets in a variety of areas.  Amongst other projects, I worked on mathematical models predicting soil quality from satellite imagery, predicting customer churn in the auto insurance industry, and predicting which job ads a job seeker would select.  It gave me a chance to show what I could do with data that I wouldn’t have been able to access otherwise.  I won prizes twice in Kaggle competitions and have been ranked as high as 15th in the overall Kaggle rankings.  I’d like to get back to a competition or two at some point, but moving a couple of times over the last year and setting up GridCure’s analytics group has been keeping me busy for now.

What do you find the most challenging at GridCure?

We’ve been working on quite a few different projects, so keeping up with the background research needed to do useful work on different predictive modeling problems is challenging, but working on the challenging applied mathematics problems that arise within the industry has been rewarding.

We have offices in three locations, so the team is still very distributed.  I’m trying to convince the others to move to Boulder, because of its mountains, 300 days of sun per year,  and less crazy housing prices than San Francisco!

What has been your favorite project at GridCure?

Definitely, the transformer predictive health module.  The physical and chemical maintenance tests that go into determining whether a large transformer is functioning are interesting to me.   Dissolved gas analysis (DGA) uses the gas concentrations in transformer oil to determine what types of faults may be occurring within a transformer.  Most utilities use linear rules of thumb based based on gas concentrations and ratios of gas concentrations to determine a transformer’s status.  Machine learning techniques can be used to combine DGA analysis with other maintenance records providing a more sophisticated assessment of transformer health.

Any favorite line from a movie?

The quote, “Just remember.  I was as good as any and better than most…” from the movie Gattaca was on my Kaggle profile for a while.  You’d have to watch the movie for the context.

Though it’s hard to choose one quote.  I had quite a few favorite movies and favorite quotes.  

Where would you like to go on a dream vacation?

I’ve been thinking that driving the Ring Road around Iceland would be fun.  Where else can you see volcanoes, glaciers, and the northern lights?  Maybe GridCure will find a partner in Iceland and I’ll have an excuse to go.  For a small company, GridCure has received quite a bit of interest from utilities outside North America, so there may be a number of work-related travel opportunities coming up in the future.