Thought Leaders in Big Data: Eastbanc Technologies, Chairman Wolf Ruzicka and Polina Reshetova, Head of Data Science (Part 2)

Wolf Ruzicka: We found some very interesting concepts that only the power of big data can bring, for instance, the causality of customer complaints. When a customer complains, you would assume that it is because of maybe heavy traffic or badly-maintained infrastructure. It turned out to be the marital status of the driver. That doesn’t mean that you should hire one set of drivers. You just have to accommodate the different type of personalities and backgrounds through guidelines in order to reduce customer complaints.

For that, we use Azure machine learning services and all kinds of other cognitive services. Of course, we had to spend an enormous amount of time, as usual, on cleansing the data and making sure that data is analyzable by these types of engines.

As a freebie, once you start going through these types of exercises, you discover a lot of interesting anecdotes in the data that even someone who owns that business isn’t aware of. For example, we could explain certain behaviors that we saw by concluding that they must have systematically deflated the tires in order to make the wheels a little faster so that the taxi meter clicks through the dollars a bit faster.

That resulted in an investigation and a slap on the hand for that particular service provider. This is one example that I can think of where AI-first thinking leads to unbelievable and very applicable and practical results.

Sramana Mitra: Talk about what IP you have put your finger on in doing this work.

Wolf Ruzicka: When you deal with very large data volumes, you need to be able to, very nimbly, move data from on-premise systems into a cloud provider. You need to be able to expose that data to your internal software developers or even to third-party developers and scientists.

A first set of IP that we developed is an API management system that is cloud-first and that provides one coherent system for APIs to be available fast. At the time of an onslaught of API calls and demand, you need to be able to move data very fast. You need to be able to understand proxy servers and multi-cloud environments very quickly. That led to the creation of a startup company called Apiphany. It was a wonderful experience for us.

The next set of IP that we incorporated was a company called Kublr. We used all of the most modern technologies from Kubernetes. We used Spinnakr from Netflix. We packaged all of those up and created significant IP around making these types of technologies enterprise-ready. That led to the creation of Kublr.

Sramana Mitra: Do you have any other examples of customer scenarios where you have done an AI-first strategy and created IP from that?

Polina Reshetova: We work with text data in unstructured text. It is my favorite topic to work on. They have a lot of unstructured text and data. Their customers are customer service companies. They have lots of complaints.

I like to work with text data mainly because many methods have been developed and can be implemented. You can apply different methods rather quickly. That helps tackle one question from different points and have this holistic approach to do text analysis for particular questions. We ended up with a comprehensive AI solution.

The goal was to minimize the amount of time we have to actually read the text to understand what’s going on. We use different approaches to get information from unstructured text and have it condensed. We hope to use this to speed up replies for each complaint and have a bit of customer service. That’s one example.

Wolf Ruzicka: The second example is the oldest startup that I’m aware of. They’re about 13 years old. They recently went into a software product direction. They’re entrenched in a certain niche which is supply chain risk management at extreme scale.

Polina Reshetova: The goal is to find as many relationships as possible between companies. We have access to an enormous amount of business text and news articles that are dated and structured. We built an AI solution to try to squeeze as much information as possible and at the end, have information on a graph. That was an interesting project for us.

This segment is part 2 in the series : Thought Leaders in Big Data: Eastbanc Technologies, Chairman Wolf Ruzicka and Polina Reshetova, Head of Data Science
1 2 3 4 5