Pioneering the fight against bot fraud with Unbotify: Data Scientist Ran Arieli’s Employee Story

Samuel Harries

Apr 12, 2019

Across every vertical, apps are losing a significant portion of their revenues to malicious in-app bots. This is why Unbotify – an Adjust company – is devoted to fighting bot fraud with an individually tailored solution for each app. Following Adjust’s acquisition of Unbotify earlier this year, it was recently released as a standalone product. To find out more about Unbotify’s groundbreaking work, we sat with Data Scientist, Ran Arieli, to discuss working life, motivations and what it takes to be a part of the Unbotify team.

Let’s start by talking about your background. How did you get started at Unbotify, and what traits have helped you succeed?

I started at Unbotify three years ago when I was still a student. I worked part-time for Unbotify for six months while I finished my studies in computer science and mathematics at Ben-Gurion University of the Negev. During my studies I also founded a charity called Selek that helps families in need. I was spending a lot of my time building the organization, which is the main reason I didn’t start working for Unbotify earlier. I was able to complete this project — which still exists in Beersheba to this day — before starting work at Unbotify full time.

Back then, Unbotify was a smaller company, with a group of employees who were friendly, extremely talented and had really big ideas. Unbotify has developed a lot since then, but the core traits needed remain the same. For example, right now we’re looking for a Senior Data Scientist. Like the rest of our team, they need to be enthusiastic about contributing to our industry-leading technology. The official requirements include experience in machine learning and big data, but we also need a really good thinker with excellent teamwork skills.

After three years with the company, what does your average day as a Data Scientist look like?

First of all, there’s really no such thing as an average day at Unbotify. Our research — and the resulting product — is unprecedented, so our day-to-day job involves solving problems that are either completely new, or that no one else has managed to solve. A great example of this is detecting click farms. Each visitor looks like a legitimate human, which means that bots have become extremely sophisticated. This is just one good example of what we’re fighting and why it’s a big challenge for the whole team.

Of course, we now have a lot of experience in fighting bots. We have the right data and have developed powerful bot detection tools. Each client is different, each use case is different, and therefore each project is different. We always plan ahead, but with new projects coming in, it’s always a super interesting challenge.

So what would you say are some of the key challenges you’re facing as a data scientist?

Over the last three years, I’ve advanced from Software Engineer to Data Engineer, and now I’m a Data Scientist. There have been a lot of challenges during this growth because what we do is pioneering in many ways. There are no open sources online or anything like that, so it’s unconventional work, and when you do overcome a challenge it’s actually a breakthrough for our field. Fraudsters are moving fast, so you need to keep up the pace and detect the most advanced bots.

However, the landscape is changing all the time, so we keep developing cutting-edge technology. As a Data Scientist, working at Unbotify is really interesting and the end results are incredibly satisfying.

We have a great team, that helps a lot. Everybody is enthusiastic and there are always people around to brainstorm with. It’s a lot of fun to take a problem that nobody has solved and think collectively to raise possible solutions. Then there are ninja engineers who will help you build a data pipeline, or whatever tool you need in order to help you manipulate it and visualize it. So every project is a big challenge but there is strong support from Unbotify’s team to help investigate, research, analyze and do whatever else is necessary to dig into the data, detect patterns and find a solution.

As part of a company with pioneering bot protection, what has been your greatest personal achievement at Unbotify so far?

There are a few things that come to mind, but one personal achievement was building Unbotify’s Offline System, our first big data detection pipeline. I helped significantly improve Unbotify’s solution, specifically the accuracy of our bot detection models. Our high accuracy rates are something we’re all proud of.

I also built the first bot database, which was one of my first projects. We needed a big database of bots in order to learn from them. We used it to understand bots so we can detect them in the real world. Thanks to that process, we were able to detect some pretty advanced bots that were imitating humans. It was this that allowed me to gain a deeper understanding of how bots work, what their capabilities are, and ultimately how this can be used to protect apps from them. Once you understand your enemy, you can detect it.

You seem to love working at Unbotify, do you have a favorite aspect of your job?

My favorite aspect of working at Unbotify is the excitement, the adrenaline, my true faith in Unbotify’s team and the product we’ve created. It’s a combination of excitement about solving broad-scale, difficult problems from our relatively small office. The team always makes it enjoyable, fun and even more interesting by sharing ideas together. Also knowing that this is an important product, and knowing that nobody will hold you back, is a wonderful feeling. My own advancement in Unbotify was only possible because Unbotify’s management enabled it.

If you’re up for it, you can always learn, advance your experience in new fields, and challenge yourself. It’s a cliché, but the sky really is the limit. This is why I’m still here, it’s full on and the future is full of opportunities. I also still manage to do charitable work for an organization run by members of my family, which is an important part of my life outside of Unbotify.

What else keeps you motivated each day?

You don’t know exactly how every day is going to end. You’re facing a problem that’s causing the industry a lot of issues, and you have really large, big data sets in front of you, tools you’ve already built and a lot of ideas — but there are still a lot of unsolved problems. The adrenaline is definitely a reason to get up and go to work.

So what would you say to someone thinking of applying for an open position?

I would say that if you’re hungry, devoted and want to be part of revolutionary, cutting-edge technology — or if you’re excited by theoretical research and a small start-up that’s confronting huge-scale fraud — Unbotify is the right place for you.

We’re currently hiring for multiple positions, but specifically for my team, you need a solid background in data science and experience in machine learning – and dealing with fraud is a big plus. This is the official part, but when it comes to personal characteristics, teamwork is a huge part of the job. We’re working together all the time. Of course, we’re also looking for nice people who are as motivated and enthusiastic as the current Unbotify team.

Looking at the future of Unbotify, what excites you most about becoming an Adjust company?

We see a lot of really interesting challenges in the foreseeable future, and being part of Adjust is really exciting. Now is a pivotal time to be joining Unbotify. We’re growing a lot in the coming year, so we’re hiring a lot and facing new challenges. We’ll see exponential growth, new projects, new challenges — so it’s a very dynamic time. We’ll continue to develop models that are robust enough to manage a lot of different use cases and different customers. So these are big times for Unbotify and Adjust.

You can find more employee stories over at our Careers page, a curated page sharing the unique experience of working at Adjust. If you’re interested in joining one of our teams, you’ll find all of our current openings here.

Want to get the latest from Adjust?