Data Science at Everstream Analytics: Meet Jack Krueger

In our second “A Day in the Life” interview, we talk to Jack Krueger. He is part of the data science team at Everstream Analytics.

He leads a key project behind our Global Monitoring and Alerting solution and shares his background, work, and love of chess.

Originally from North Carolina, Jack moved to Germany for a master’s degree in social and economic data science. His thesis took him in the direction of natural language processing focusing on bias in the media.

This background made him the perfect fit to join the Everstream Analytics data science team that focuses on what is known as the “Daily Planet AI Pipeline.”

“It was a big advantage that I had worked with news media, and I had worked with natural language processing. So, it was kind of a natural fit for joining the team to build the Daily Planet AI Pipeline.

“It just felt exciting, doing something that was at the cutting edge, doing something that was pushing the limits of what’s possible, what’s new. That is always motivating,” he says.

But what is Daily Planet AI Pipeline, and how does it work?

The Daily Planet AI Pipeline

The Daily Planet AI Pipeline feeds the Everstream Analytics Global Monitoring and Alerting solution. This solution uses artificial intelligence to detect supply chain disruptions.

“We start by scraping as much news as we can from the internet constantly,” says Jack.

Incoming data goes through a series of classifiers, algorithms, and filters. This removes any news that is unrelated to supply chains. A strict set of guidelines determines what is newsworthy, reportable, and relevant.

“As a news story moves down the pipeline, we start to enrich it. What is the topic? Is this about an industrial fire or a port stoppage? What are the companies mentioned, the products mentioned, the people mentioned?”

Many different news sources cover major events. These are written in several different languages. Algorithms detect the language and decide whether they need to translate it. In addition, an algorithm consolidates news sources about the same event.

“Our algorithm figures out if this is an article about an event we are already covering. Is it part of the same story, just written by a different publication? Does it provide follow-on detail? If so, we combine them so that we keep all the relevant information without creating redundant events.”

At the final stage, the algorithm summarizes the articles and delivers this draft, along with clustered articles, to the analysts in the Everstream Analytics Intelligence Solutions team.

AI-Powered, Human Validated Alerts

The Intelligence Solutions team is Jack’s most important internal customer. This global team of analysts provide human oversight before the platform receives an alert.

In addition, the analysts provide regular reports on a number of topics impacting supply chains. These range from geopolitical relations, protests, port congestion, natural disasters, and weather events.

“We provide this curated newsfeed for them that makes it as easy as possible for them to do their jobs. They log in to our newsfeed and can easily see what’s most relevant,” explains Jack.

Continuous Improvement

Jack joined the company in early 2022. In November of that year, OpenAI released ChatGPT and stunned the world with its ability to generate text, human-like answers, and perform various tasks.

The data science team was quick to see the possibilities of generative AI for their work.

“Because the rise of generative AI came in the field of natural language processing, because the most advanced AI is so heavily language based, it became really natural from the beginning for the team that’s working on news to be the team that’s taking advantage of the hottest and newest stuff.

“The first thing that we did was use generative AI to analyze and summarize articles, then format them in a way that we like our incident reports to be formatted,” he says.

The Everstream Analytics team now leverages multiple AI agents, filters, and algorithms across several data science projects. In addition, Jack and his team are always looking at new ways AI can be used to improve the Daily Planet AI Pipeline. It is a good thing that he cites “problem solving” as his favorite part of his job.

“I like problem solving. I think that’s one of the coolest parts of data science.

“So, if you’re building a new classifier or a new model, it’s really a lot about thinking about: Okay, how do I set this up? What data do I need? How do I set up the data that I need? What problem am I trying to solve? How do I go about trying to solve it? And that’s a really fun process.”

Day-to-Day Life

Jack is currently living in Spain and has a routine he enjoys.

“I have most of my meetings on East Coast mornings. Starting off from maybe 2pm my time. That means that my mornings are quieter. It’s time for me to do head down work on my own.

“I have a co-working space here. Oftentimes I’ll go there in the mornings, get my work done and come back to my apartment for lunch, then have my meetings in the afternoon.

“It’s a nice rhythm where I check in with everyone in the afternoons, figure out what needs to be done, and update people on what I’m working on. I take notes at the end of the day on everything we talked about that afternoon. Then I have time on my own the next morning to get it done. It is a different rhythm, but it’s a rhythm that I’ve grown to appreciate.”

Outside of work, Jack has a passion for chess.

“I play a lot! One of the first things I did when moving here was join a local chess club to try and meet people.

“I go almost every Tuesday to my chess club. I compete in a league here on Sundays in the wintertime. It has been very important to me, in my life, and in moving here,” he explains.