Swaroop Jagadish & Shirshanka Das, DataHub Co-Founders on Data Infra

**Grace Gong** (0:00)
I think I clicked, we're live, sorry. Hi, Swaroop and Shirshanka, welcome to Venture with Grace.

**Swaroop Jagadish** (0:10)
Great to be here, Grace.

**Grace Gong** (0:13)
I'm sorry.

**Shirshanka Das** (0:14)
Amazing to be here, and I'm sure we will get better at figuring out who speaks right after whom.

**Grace Gong** (0:20)
Oh my God, I absolutely love, like, you know, this is not the first time like we have two co-founders, we had the founders of, yeah, we have like quite a few, like two people on, I absolutely love the two co-founder format because there's always like, the story is always slightly different from two co-founders. So I want to start with a quick shout out to our amazing sponsor. This episode is brought to you by Nebius, the ultimate club for AI innovators. Nebius provides AI infrastructure you can count on, commanding reliability and speed with flexibility, and engineering support on match by hyperscalers. AI leaders like Meta, Shopify and Higgsfield already partnered with Nebius to run their AI workloads. Plus, Venture back startup can save up to 150K on compute cost when they apply for access. Visit nebius.com or nebius.com/startup to learn more, or use code grace you may get a surprise this time.
I want to start with your journey. Both of you actually come from big company background. So, Swaroop was at Airbnb and LinkedIn, Yahoo, and then both of you, I believe, work at LinkedIn, Yahoo.
So, maybe we could start from what were some early career lessons that you've learned early on that impacted you later on as a founder?

**Shirshanka Das** (1:40)
Well, I'd say there's a couple of lessons that we learned, I personally learned in the early days of my career that kind of ended up getting turned around its head later on, which I think is pretty interesting. When I started at, I would say my career in data really started at LinkedIn. Before that, I spent a lot of my time in distributed data systems, et cetera, distributed systems at Yahoo and PayPal. But LinkedIn is where I started really working with data and started out in the online data infrastructure space, and then moved to streaming, and then finally batch large data or big data as it was called back in the day.
One of the things that was repeatedly a lesson that I kept on learning in my career was that as data gets transformed and as innovation happens, you have to keep first doing the first thing that you need to solve for and then you move to the next. So kind of that sequential process of you don't just get to the right answer until you've gotten your housing order was something that I learned early on.
Essentially, anytime you encounter chaos, you first have to organize your data, you then have to annotate it, you have to make it discoverable, understandable, and then you can start using it. And the same thing was true even for technology in many ways.
The first time we encountered the idea of scalable storage. I said, okay, all innovation projects are on pause. We'll first move to scalable storage. And at that time, it was HDFS for us. And then we will start unlocking all this amazing data that LinkedIn has. So there was a very strong push towards adoption of any new technology was gated behind adoption of the previous foundational shift in data. And I thought that that was the lesson that I had learned at LinkedIn. But more recently, what I'm finding is that with the advent of AI, a lot of those sequential transformation journeys are getting disrupted.
AI adoption and the speed of AI adoption has actually been pretty amazing to watch. We now regularly encounter customers and community members who have a messy data ecosystem, but they're not waiting for the full migration to the cloud, or they're not waiting for getting their house fully in order. AI is showing up, people are connecting stuff to their warehouse or to even online systems, like even physics is getting violated now, right? Previously, the whole theory was analytical data is where you go to ask questions. But now people are just connecting cloud up to whatever they can get their hands on. If it has an MCP, then it's a warm body I can connect to and get answers, right? So a lot of those lessons that we learned, or I learned at least have been turned on their head, because AI has been such a disrupter force that people are not waiting to get started with processing their data, transforming their data, making sense of their data, and doing other interesting things with their data. So I think it's a pretty interesting moment where we are changing the way in which things were done in the past.