Gideon Rubin: Then once they had that estimate and were comfortable moving ahead, then we pull in the data and every day, we process. It changes the bids across Google and other channels. It updates the budgets. It keeps reprocessing. We allow them to access the data. We allow them to automate the movement of the data. At the end, we allow them to get insights that enable their clients to be highly successful in their marketing campaign.
With that new positioning, they went public in August pushing machine learning as a core competency. For that 140-team that I mentioned, the product is 500% larger as far as customer base now than what it was when we started. They cut down from six data scientists to four. Their DevOps were cut drastically. That’s just one example.
Sramana Mitra: Do you operate as a services company? Do you take these projects and apply your expertise to achieve the kinds of outcomes you’re talking about?
Gideon Rubin: We’re a SaaS company. Large-scale businesses can go to our website. We have three main solutions. They can all be tried for free. Pandio ML is our machine learning Python library. That’s open-sourced. We allow them to access disparate data sources through a central interface.
The third piece is about data logistics. Think about distributed messaging. Companies can try all those for free. When they put them all together, that’s what we call AI orchestration. That allows them to orchestrate all these pieces to one central place.
Sramana Mitra: What kind of usage do you have for the open source piece of this?
Josh Odmark: We just launched that a few weeks ago. We got very good feedback. Several people are using it. Our focus is primarily large-scale enterprises. Those are the ones with the most challenging data problems. Our capabilities, vis-à-vis performance and scale, are far beyond every solution in the market. That’s where you see massive cost and time savings. With that, I would say that several of the top financial institutions in the world have started using it. Airlines have started using it. Many large media companies too.
Sramana Mitra: Talk to me a bit about the data infrastructure that you interface with. What role do the NoSQL databases play? Where do you sit? Where do they sit?
Josh Odmark: The first piece of accessibility technology that Gideon mentioned is a query-in-place engine which lets you run SQL against any sort of data source. In that case, we’re an enabler of those systems. If you’ve got data in Postgres, MySQL, Dynamo, or Snowflake, very rarely do you have all your data in one data store where you can execute a single query against that data store and have everything you need.
Usually you need to join at least two sources. In many cases, it’s 10 plus. Our query engine lets you join across different database technologies in a single query. You can also query petabytes of data with it as well. That’s incredibly powerful. We work with a few databases right now, but we support pretty much any type of protocol or database that’s available as well as real-time streaming frameworks.
You can run an SQL query against a live stream of data or join it against a data that is at rest somewhere. Its main purpose is to be a feeder into something. Maybe that powers a dashboard so that you can show some sort of metrics to your customers. It can also be a feeder into something like Spark or another system.
Sramana Mitra: Talk to me about Apache Cassandra and DataStax. How do you work with them?
Josh Odmark: They’re just another database that you can query into. In the Pandio platform, you would go and add a connection to Cassandra and you would enter your credentials to make that connection. Then you can start running queries against that database. Then you would add something else out there like Snowflake or DynamoDB. You would add those with your user access credentials. Now you can execute a single query against all of them and join different tables. The way we work with them is to make it easier for people to get data out of those systems with a single interface into multiple databases.
This segment is part 2 in the series : Thought Leaders in Artificial Intelligence: Pandio Founders Gideon Rubin, CEO and Josh Odmark, CTO
1 2 3