Thought Leaders in Artificial Intelligence: Pandio Founders Gideon Rubin, CEO and Josh Odmark, CTO (Part 3)

Sramana Mitra: Do you see Cassandra a lot?

Josh Odmark: Yes, it’s been popping up a lot more lately when it comes to machine learning because of its read and write speed. People treat it both as a traditional datastore and also an intermediary. With data science, there’s typically layers to your model or a bunch of steps you need to take until you get to your model. We see Cassandra being used both as a data store and a place to put the intermediary steps. In the streaming space, Cassandra is popular.

Sramana Mitra: Do you see startups building applications on the Cassandra database?

Josh Odmark: Yes, for sure. It’s relatively new technology, so they’re using Cassandra if it solves a very specific use case for them. Streaming is a great example. If you’re going to stream data, it’s coming at a high velocity. You need something like Cassandra that can handle it. It’s not the only solution, but it’s a popular one right now.

Gideon Rubin: When it comes to startups, we see a lot of situations where they really have great ideas and they end up using a lot of these cloud services because they’re available and scalable. They prop something up through Amazon or some other provider. But it doesn’t actually mean that it’s scalable or production-ready. Startups want to transition to a much more stable and scalable infrastructure. They don’t have to manage it themselves. That’s where we come in oftentimes.

Sramana Mitra: How many startups do you work with?

Gideon Rubin: We don’t disclose our numbers at this point, but it’s in the dozens. If you look at my background, I was a Champion for Startup America. I have pretty deep ties in the startup world, and we’re just starting to fully commercialize what we’ve built. We’re starting to engage with the startup world now. We had some great responses in the last few weeks. Some of our earliest customers were founders and some other VC-backed startups. 

Sramana Mitra: What are some open problems or startup ideas that come to you that new entrepreneurs who are going to be reading this interview should be looking at?

Josh Odmark: There’s two that I would love to be solved. One is a particularly difficult one. That’s around data sharing. One of the things that’s difficult these days is, let’s say Facebook wants to share their data with Google. That’s never going to happen. One of the struggles of machine learning and artificial intelligence is having access to data.

Right now, only the big corporations have enough access to data to do certain models. For some machine learning algorithms, you don’t need a huge amount of data. For others, you do. The more data you have, the better it gets. Data sharing is very difficult. It’s not just a technology problem. That’s my space and what gets me interested. There are other things out there like homomorphic encryption where you can encrypt the data and give the data over.

Somebody could learn from that encrypted dataset. They can’t see the raw data, but the relationships between the data points remain. You can perform math against it. There are some problems that still exist in that space. One being it’s orders of magnitude more difficult to compute against encrypted data. Microsoft is making huge strides there. We have actually talked to a number of companies who are in the space.

In order for AI to really permeate the world, someone’s got to solve the data sharing problem. That can just be making data available digitally or helping companies share data with confidence. There’s likely to be political battles there too. You can imagine all the problems trying to get Google to share with Facebook and vice versa. That’s a big one for me. Everybody wants AI, but very few people have it.

Sramana Mitra: Thank you for your time.

This segment is part 3 in the series : Thought Leaders in Artificial Intelligence: Pandio Founders Gideon Rubin, CEO and Josh Odmark, CTO
1 2 3