What Dog Days of Summer?
Get the Facts on our Incredible Sisu Interns

By Grant Shirk - August 12, 2019

Not all facts have to be serious (nor, in fact, do all blog posts). So here’s something a little different – a fun fact about summertime, and a few more insights into our awesome summer interns at Sisu.

Where do the “dog days of summer” originate?
As it turns out, it’s all about astronomy. There are multiple citations, but as early as Homer’s The Iliad, the Romans noted that Sirius, the Dog Star, was rising just before the sun, in late July.

And now, back to our interns. This summer, we welcomed Julia Belk and Leo Mehr to the Sisu team for a crash course in startups, agile process, and operational analytics. Along the way, they’ve shipped production code, spent the night on a ranch, and attended executive briefings at Andreessen Horowitz. Not a bad way to spend these dog days.

I took a moment at the end of their term as Sisu interns and sat down with Julia and Leo to talk about their backgrounds, what they learned this summer, and what surprised them most about Sisu and taking the leap into the world of enterprise analytics.

Before we get started, tell me a little bit about yourselves. Where did you grow up? What’s something surprising about you?

Julia Belk: Sure. I grew up in Texas – about an hour north of Dallas – and migrated east for college. My go-to fun fact is that I can ride a unicycle (and I was pretty into it growing up), but I don’t do it much any more. I’ve recently started bicycling more, which is essentially mandatory around the Stanford campus, and have been enjoying that a lot!

Leo Mehr: I grew up all over the US, between Texas, Massachusetts, California, and New York. My parents constantly moved as they sought new opportunities, which is originally what brought them to the US from Ukraine, where they were born. Russian was my first language and I played a lot of hockey growing up.

You’re both doing graduate work at Stanford. What are you working towards on campus?

JB: I’m a Ph.D. student in the computer science department, so I do research in a lab. I’m mostly interested in data-driven approaches to problems in biology and medicine, particularly cancer therapeutics – there has been a recent revolution in treatment thanks to “immunotherapies” that use the body’s own immune system to fight cancer. However, most patients don’t respond, many types of cancer are intractable, and the molecular reasons that underpin these differences in response are not known. From an engineering perspective, I think that if we understood a bit more about when these drugs work (and don’t work) we’d have a better chance of improving them and broadening their impact.

LM: I’m doing an MS in computer science, focusing on the intersection of machine learning and data systems. I’ve had the fortune to serve as the head TA for Stanford’s intro database course (CS 145), and to work in the natural language processing group. On the side, I’ve enjoyed getting to know many fellow students, and am currently a social chair for the MSCS student body.

What other options were you considering before you chose this internship at Sisu?

LM: I looked at opportunities big and small, from Facebook to Robinhood, but Sisu was a clear winner. Not only was I impressed by the team, but also by the chance to shape the core product and culture over the course of a single summer.

JB: For me, the main other option would be to hang out at Stanford for the summer. However, I wanted to take a break from research and learn about something completely different, and Sisu was the perfect fit! [Ed. note – Sounds like we got you to trade the Farm for the Ranch!]

Do you have a favorite project from the past two months?

LM: We actually worked together on a project during this summer’s Sisu Hackathon. Julia and I developed a machine learning model to rank facts based on users’ interactions with them. For example, every time a user looks at or favorites a fact, this powers our model to show future users facts they are more likely to find interesting. This started as a simple hackathon project, but showed very promising results and should be deployed to production soon.

JB: I really enjoyed working on that project as well. It was super fun to work as a team and prototype a riskier feature outside of the normal engineering development process. We successfully extended Sisu’s current statistical ranking and relevance processes by incorporating implicit user feedback into the model. And it’s doubly cool that we get to follow up on it and see it live in the product!

What was the most surprising thing about your summer here?

JB: Before joining, I didn’t fully realize how much the company invests in to strengthening the team, doing team activities, and spending time together outside of the standard engineering context. I feel like these activities (Reading Group lunch, the Sisu Retreat, the Sisu BBQ, Hackathon, etc) are really important long term investments in the team and were also super fun!

LM: For me, it’s that some of the work I would do could be critical to securing deals with customers. I didn’t expect to be given a project that could unblock opportunities with Fortune 500 clients in my first few weeks here. This reflects a culture of empowerment and risk-taking that I’ve seen at Sisu, qualities that make it a very exciting place to work.

Any advice for next summer’s interns?

JB: One of the unique and awesome things about Sisu is that you’ll have a voice in all aspects of the organization – not just building features but also meeting with customers, iterating on the product design, and learning/growing with the whole team. Be prepared to learn a lot and have fun!

LM: Don’t be shy to give feedback about anything, from onboarding and documentation to organizational process and cultural norms. Brush up on React, spend time getting to know your coworkers, and ultimately, focus on learning and having fun!

Julia and Leo, thanks for taking the time to answer these questions and share a bit of your experiences with us. You’ve both had a big impact on Sisu (as a company and a product) and each of you have delivered some serious WOW from the moment you first stepped through our doors. Glad to hear that fun still managed to be part of the equation, too.

Good luck with your next adventures!

Read more

Towards Off-the-Shelf Machine Learning

Achieving high accuracy in machine learning often requires extremely large amounts of training data – but just how much training data is required?

Read more

What to do with Big Data?
Making ML useful is a platform problem

With today’s excitement about making AI and machine learning useful, it’s easy to forget that we were only recently enamored with Big Data and its promise.

Read more