Thought Leaders in Big Data: Interview with Brian Bulkowski and Srini Srinivasan, Co-founders of Aerospike (Part 6)

Sramana Mitra: How big is the Bangalore team?

Srini Srinivasan: We have about 25 people.

SM: Did you lose people in Bangalore?

SS: We have been very fortunate to have very little attrition. Part of it is because of the people we attract wanting to work in cutting-edge technologies. You don’t give these kinds of jobs very much, even in the Valley.

SM: So where do you hire from?

SS: Essentially people found us, because we have a person – also a founder – who runs our team in Bangalore. He figured out a person through his network who is very solid in database terms. If you want to build a database from scratch, there are few people who are crazy enough to want to do it. It is a hard problem, and it only happens once in 30 years. Then you find them and you keep them.

BB: Nothing else excites them. Another way to think of it is that there is a whole way of how the IBM DB2 group went to Bangalore and the Sybase group went to Pune. They do their core engineering, and they trained a whole generation of performance-oriented C programmers. That is what we need, which is why we have around 40 people. I think of them as the TRS80 cohort – people who were brought up with DOS and Apple 2. The parallel in India is that the “30-somethings” who were there when Northern Telecom came in and said “We are going to train thousands of C programmers.” It is not about price, it is about the quality of people.

SS: We have an entire team there, with every department of engineering and operations represented. That is our philosophy of setting up distributed teams – having every piece of it locally. The basic thing is that a team should be able to fully function in its time zone. If it can’t, it will not work. On the other hand, you are going to have meetings with people in other time zones. That is why we have two teams. One time Brian and I were looking at each other when we were doing something at night, and we said, “Why are we both in the same time zone and we are doing this? The first thing we do when we get money is we are setting up a team in Bangalore.” Even before the money hit the bank, I was out finding a team in Bangalore. Then I got some sleep back.

SM: What is happening in the industry? What are the open problems?

BB: I think we are only partly through this database revolution. Sitting where we are, the world looks like  NoSQL is a done deal. Then there are parts of Manhattan that also feel this way. Then you get outside of that and you see a different landscape, where this revolution of orders of magnitude haven’t occurred yet. This revolution has at least a decade to play out. I think a lot of that has to do with the polyglot nature of development – the different languages and language formats. I can use Scala to do my framework, but I still need to use SQL? Why don’t I also have a polyglot database underneath me? That is one of the problems I think we are right in the middle of in this shift. Part of it is that the big data revolution is already upon us. It is on the cover of the Harvard Business Review. However, what do you do with the data? You get these big analysis, what do you do with it? That is the problem we are trying to address now. But it does have to do with new applications, new ways of programming, faster development times, etc. The tools of the past are not falling by the wayside, but they are being added to.

SS: As we see on the adoption side, most of the new database technologies have come out of Internet enterprises. That is where the biggest needs for interactions occur. For an Internet enterprise to be successful completely depends on holding the attention of their customer. I think all Internet enterprises are going to have to move to that model over time if they want to be successful. If you look at the largest enterprises in the world, a lot of them are Internet enterprises: Apple, Google, etc.

Given that a lot of enterprises have to move, many of the problems they have today – they probably have reports once a week or once a day – all of that can actually be done in real time, which means they can compete effectively. Some companies like Walmart have actually figured it out. For example, how do you take a number of smaller retailers and make sure they somehow benefit from the Internet profile of the user that is walking into their store? Walmart or Amazon are big enough to have it, but what about a smaller store? How do they do that? There is a huge opportunity there. That is why you need systems to work in scale.

SM: But that is not enough. You guys provide the infrastructure. Who provides the application on top of it to be able to feed into these little e-commerce retailers? They can’t do that by themselves.

BB: There are several components to it. One is where do they actually get the data from?

This segment is part 6 in the series : Thought Leaders in Big Data: Interview with Brian Bulkowski and Srini Srinivasan, Co-founders of Aerospike
1 2 3 4 5 6 7