Skip to main content

Summit Live: Fireside Chat with Arsalan Tavakoli, Databricks co-Founder

Summary

  • Leadership perspective: Arsalan Tavakoli, Databricks co-founder, shares insights on implementing and scaling data intelligence across enterprises.
  • Field engineering approach: Databricks takes a hands-on, integrated approach to customer success by breaking down traditional silos between pre-sales and post-sales functions.
  • Community growth: The Databricks Data and AI Summit has grown to attract over 65,000 global attendees with watch parties worldwide.

Summit Live: Fireside Chat with Arsalan Tavakoli, Databricks co-Founder

Watch: Summit Live: Fireside Chat with Arsalan Tavakoli, Databricks co-Founder
Arsalan Tavakoli-Shiraji co-founded Databricks, growing it to one of the most influential tech companies in history. Arsalan will expand from the mainstage keynotes, and discuss how companies are implementing, monetizing, and scaling data intelligence.

Chapters

FAQs

What does field engineering mean at Databricks?

Field engineering at Databricks represents an integrated approach to customer success that breaks down traditional silos between pre-sales and post-sales, allowing technical teams to work hands-on with customers to ensure their success with the data platform.

How has Databricks grown since its founding?

Databricks has evolved from a small company starting with Spark Summit to now hosting a major summit with over 65,000 attendees from around the world, with watch parties in multiple continents including London, Amsterdam, Tokyo, Brazil, Costa Rica, and India.

What are the key challenges companies face implementing data intelligence?

According to this video, companies need guidance on how to implement, monetize, and scale data intelligence solutions, which requires taking technical capabilities and connecting them to real business outcomes and customer problems.

How does Databricks approach customer success?

Databricks emphasizes a hands-on approach to customer success in the technical data space, believing that successful customer outcomes require direct engagement and integration across different aspects of the business.

Full transcript

0:00 Intro

All right, we are back live summit live here at the data bricks data and AI summit. Ari Kaplan, head of evangelism at data bricks. Hi. Sorry you caught me taking some very serious notes here. Yeah, one of the things we're talking about, we have had an exciting morning with incredible announcements. We're about to bring Aralon on and we and I have an exciting lineup coming up in the next couple hours. Oh yeah. We're talking data platform.
We're talking everyone's favorite thing, data migrations and then we're hopping into AI for marketing. So we've got a hot lineup coming. Great. And yeah, we have Arselon coming on now. Welcome. Welcome. Thanks for taking the time. Thanks for having me. Appreciate it. See you. How are you? Amazing. We have uh first of all 65,000 people plus tuning in from all over the world. Watch parties. They uh sending sending in their images.

1:06 Audience

Theaters in London, theaters in Amsterdam, Tokyo, Brazil, India. So, it's amazing. The numbers are still a little crazy, right? I still remember when we were a lot smaller. We were not throwing around figures like 60,000 plus. So, incredible to see. Well, I have a very hard-hitting question for you. I want to start with if you were a professional wrestler, what would your walk-on song be? Oh, uh, Guns and Roses, sweet child of mine.
Nice. You knew it. Slash 100%. You know, called that. Yeah. Now I want to know how knew that, you know, like. So, but intuition. You have that incredible beyond life persona. Yeah. Yeah. Congratulations on all the success and building it. I know it wasn't just you wake up and it's a major company now, you know, starting small with that Spark Summit and building it out. Um, now so you've had so many different experiences in life.

1:38 Intuition

Field engineering, the team that you lead, you're on the field engineering. I am. So I kind of wanted to ask like what is field engineering? What is field engineering? Right. Well, look, I think field engineering there's the diplomatic answer and then the actual one. Um in a lot of other organizations it takes a lot of work to make customers successful ultimately and they break them up into a bunch of different silos. is they have very traditional names like oh this is pre-sales this is post sales this is special for us we look at it there's no question um data bricks and data and AI is a technical space just in general right and one of our ethos has always been we want to be hands-on with customers making them successful and we want to bring all those resources to one place not just ship an org chart to customers so field engineering is basically anything technical that does not sit inside of core R&D everything from how do you help a customer understand exactly what this is to how do you make them successful going live and how do you train and develop their you know people so you know it's a testament to the investment of data bricks I think about roughly a third of the company sits inside of field engineering so lots of investment making sure customers are success successful that's great I know there's like multiple hours of announcements heard tons of rounds of applause incredible things and not everyone uh you know has heard everything but I want to know what like what were some of the most exciting parts from this morning's keynote.

2:20 Field Engineering

Oh, this is dangerous. This is like when they asked me to pick a favorite child. I've learned how to deflect that. Uh, you know, at this point, um, look, I think that there was a bunch of really interesting things, but for me, two, you know, I'll give you a couple that, uh, stood out, so can't just pick one. Obviously, lakebas is a new category is incredibly interesting. I mean, sitting with customers, the ones who are just like, we do analytics, but now it's so intertwined with how do you take applications into production?

3:39 Most exciting parts of the keynote

The fact of having Lake Base in a database that's available now I think is incredible. Um I think agent bricks this notion that everybody's been trying to figure out yes I need to use AI how do I exactly do it how do I go about it and I think agent bricks is what I've seen is the biggest step forward in how we think people will actually engage and leverage it without necessarily needing to get into all the underlying details.
And the other one I would be remiss not to mention is I think uh data bricks is uh free edition. Yes. Big deal. Big deal. Right. We and not just because we come from academic roots at, but it's always been this huge investment. How do we uplevel everybody? How do we make sure that people can learn and so that everybody now can get their own environment. Open sourcing kind of all of data bricks academy and that content for everybody to be able to become an AI expert, a data expert is super exciting.
Yeah. Yeah, the free edition and then also a significant investment into growing that learning community. Absolutely 100%. I know people are so excited about the free edition, myself included, because I have people asking me, you know, how do I just try this out? How do I get my feet wet a little bit? And it was always a difficult question. Now it's so simple, you know, send them in, jump in. Exactly. Exactly. No, I think there's a bunch of people who already said this is great.

5:10 How do I get my feet wet

No credit card, sign up, get going, and get the full power, you know, of data bricks, being able to use the assistant, all of those capabilities. I think we've seen a huge amount of demand. And I just think universities, everybody, even on a personal basis, everybody's telling you in this data and AI era, you need to learn, you need to be hands-on and understand it um to be able to move forward. So, giving them a vehicle to do that and the content for it.
I think we're excited to be able to do so. Yeah. And want to hear like maybe one of your exciting news for me, you know, Lake Base and OLTP. I come I used to be the head of the worldwide Oracle user group here in Moscone. And like that's one of the pieces companies have um uh structured unstructured streaming uh you know OLTP data warehousing and to have that unifying which will be all over unity catalog with the governance just makes a ton of sense.
So I'm personally excited in that and just how much the market and the value that customers will get bringing that into that larger unified umbrella. Yeah, I was a data bricks customer less than a year ago and I remember asking when are we getting OLTP my team needs it like soon like I don't know if I believe that and now you know hey we we try we try to underpromise and overd deliver not the not not the inverse I think best way the exciting part about an announcement like that is you hear it and everybody's like okay great well must have just done an acquisition god knows when I'm going to actually get this and you can get it like here's the hundreds of people who've already been using and it's available today go get it start using I think being able to say that I think also gives people a lot more confidence that it's not just you know marketing uh slides but something that they can actually use.

6:22 Underpromise overdeliver

It's not a beta test. It's ready to rock. Ready to rock and roll. And speaking of rock and rolling like the data bricks apps like that uptrend like just even the public preview to you know thousands of customers applications like really the fastest growing in data bricks history. And we're going to have come on you know go beyond what was said earlier on the keynote. So that's you know another just don't tell him that when I picked three and apps wasn't one of them cuz I love apps cuz then he'll be like then you're not a favorite child.
Just don't mention that to him. But there's too much too much to go around. Exactly. Yeah. And I have to ask so started this company. What would you go back in time and tell baby Arselon? Like what what do you wish you knew maybe five 10 years ago? Uh, look, I I think it's always a hard question because you look at it's like the, you know, the butterfly effect. If you change something to the kind of affect where you end up.

7:30 What do you wish you knew

I think the most interesting thing for us was there was a bunch of, I think, decisions we made that you sweated a lot. You knew that they were the right decision, but you sweated a lot about them. And at the end of the day, you look back and you're like, just stick to your guns on it. Like, I'll give you an example, a big one. I cannot tell you how much pressure there was in the early days that said just go on prem. All the money is on on prem.
Nobody's going to use cloud. Nothing will be like literally people told us why would you want to focus a company on places where nobody will ever actually put real data, right? And so um you know so the fact that you stuck to your guns there I think one was a really you know was a really big deal. The second part that's also interesting is uh it's always good to be paranoid but don't sweat the small details. Like I still remember we would um in the early days you were focused on trying to build the company but you were tiny.
You were trying to figure out how to not die right at that point. And these companies that looked around you that look like behemoths and the like about them and you're like oh my god did you see what they put out in a press release? What did they do? What does that mean for you? I think you probably spent more energy there versus just being like look there's a lot of es and flows in this industry just focus all your energy on how do you if you have a vision how do you execute on that do right by customers and I think that was overwhelmingly what we did but there's definitely times where you lost a little bit of sleep wondering what this or this would announce so that notion of you've got a path stick to it execute and if you do that well if you had a great vision and execute you'll be in good shape I think those would be kind of I think more comforting and that's probably what comes from more experienced entrepreneurs than we knew that they kind of knew what that journey uh would look like.
Yeah. Follow your intuition. Maybe don't give into the FOMO if you're looking around. Don't follow everyone else. Right. Blaze your own path. Exactly. Exactly. Now, one of the things we we love our virtual audience coming here in person. So many uh great reasons. And you've had the last couple of days, I know, executive form customer meetings telling people, you know, how can you accelerate your AI journey? I mean, speaking of acceleration, Naveen Ralph can't be here, but in my eyesight is that uh his car like he is like and that's actually how I met him beforehand doing and Drey consulting with them.
They're like, "You got to meet this person before like we both ended up here. He's a race car driver." So, uh and the one of the fun play on words is acceleration using his car as like a prop. But that's re realistically one of the themes of what you've been talking about, what we've all been talking about is how data and AI, how can it accelerate customer journeys. So just want to hear like some key pieces of advice you've been giving to these forums the last couple days.
Yeah. Well, well look, I think if you were to go back a couple of years ago, the question was much more or still oftentimes like, okay, well, why would I pick data bricks versus X or Y? I think that doesn't really happen as much today. I don't mean that in an arrogant way, but I think like people are kind of sold on the power of the platform and what does it look like? I think now people are saying great, help me understand how do I get there fast, right?

10:23 Key pieces of advice

Because you got to remember most people's journey with any platform, data bricks is no different, is it starts with a migration? And so they're looking at do I lift and shift? How do I make sure I follow the best practices? So one of the big things we've tried to do is say here's what are the blueprints for success. You always focus on the technology. Don't forget about the people developing them. Don't forget about the processes there.
And then how do you how do we get involved to make sure um what you do? You pick you you basically put things in place in the right way so you don't end up saying, "Okay, I got to the end state. I had no idea you were going to announce all of these things that I need to do differently." So that notion of knowing here's where you want to end up. How do we get you there faster? There's some exciting announcements tomorrow as well around how do we help people like kind of accelerate their migrations?
But I think that's the biggest one. the faster we help them get there, the sooner they realize the benefits and they can start leveraging many of the new capabilities, you know, that we've talked about today and before. Time to value, it's always always the thing you got to figure out, right? Yeah. And this has been an exciting year. We raised a bunch of money, uh, added a ton of value to customers, but also had some key acquisitions, you know, uh, you know, pre-announced Neon not too long ago.
And don't ask me to pick my favorite one. Well, the good thing is there aren't that many that you could probably mention all of them. uh you know and even not at you know the partnership with SAP yeah is huge I I've been helping enabling with that and you know there's so much green field out there so like uh like what are some of the excitements of expanding like the data intelligence platform yeah look I think it's uh there's always more to go but I think that the really exciting part for us is to say you go back to what the company was founded 12 years ago a bit at this Right?
You had this broad vision, but the part of it that you delivered on was this kind of narrow slice, right? You you really were, you know, you were a data science notebook company, you know, in the very early days that also provided spark and some data engineering. And we've been on this journey to say, what else do people need to really kind of unlock insights? Um, and as you look at that journey, you said, okay, I have that piece.
Do I add data warehousing? Do I add governance? Um, we always kind of had the machine learning, but do I add visualization, AI, BI, make ingest better, some of the big things that we continue to get excited about is one, and I won't kind of like steal the thunder, but even more, you know, we'll talk about tomorrow, how do you make it easy to get data from all of the places that you care about with partnerships like SAP and the like.
Um, now to the other side of how are we actually going to um build true applications, right? like the OOLTP one was a big gap that people would have to figure out how do they do separately and then you take a look at that and say if I can get all the data in and I can build applications and I have things like agent bricks talked a little bit about it I think that's where the future of applications go so now you start looking at what are the domain of kind of different types of vertical specific applications that customers can build we can work with partners and you know kind of whether SI and ISV I think that's that notion where you truly can say here is an end-to-end vision of how people make data and AI work.
I think that's the most exciting part for us of finally getting to that vision and there's always going to be more that you can add. Uh but we're excited to see the types of things we can help customers do. Yeah, absolutely. And you know, one of the other things, yeah, industry by industry. Yeah, it the vertical and the horizontal both add specific uh value uh solution accelerators now even focusing on the marketing persona.
Yeah. Data intelligence for marketing. We'll be expanding more on that uh throughout the day. So exciting. All right. I'm sorry. I have no chill. I have to ask you another hard question. I want to know what is your unpopular opinion about AI? What is my unpopular uh that uh all the value will be in boring AI? I think everybody is like everybody gets really excited. I mean it's awesome when I see like the robot arm serving you here.

14:32 Boring AI

Like it's very cool or like the humanoid doing X. But I think the challenge is like those are the interesting parts but really where most of the value I think in AI will be is taking many of the things that we do every day today and you will not necessarily look at it if you're not trained and say that's AI look at the I'm sure they use agent bricks for that but it will be how do we take things that we do better we're already seeing that of whether it's generating emails or how we interact with banks or how we basically purchase different B pieces how do we take loans out so I think a lot of that is it's embedded of improving those things and bringing out new capab capabilities.
I think everybody always likes like sexy headlines or like sexy use cases are where they get it. For me, the boring pieces are that's where I think AI is really it's where there's a lot of spend today and it's where there's a lot of challenges. So, I think that notion of focus on those problems less on what's the newest model and what's the newest cool use case that uh is on TV. I usually get shouted down because I'm like, "Oh, he's a boring guy."
But I'm like, I spend the time with the customers. That's what they care about, right? There's very few customers I can go to and say, "Would you like a robot arm that can serve kids?" Exactly. It sounds cool. You might have a demo on these times, but it's not really what drives the business. And you can see a data bricks robot walking around. I don't know if you saw that or not yet. I did not. It wouldn't surprise me.
You know, I think this conference has gotten big. It used to be I could walk the whole see everything. Now like Arcelon, did you see that? I was like, I'm sorry I didn't get a chance, but I know there's a bunch of good I feel like I spend the week or two afterwards looking through all of the footage of hearing, "Oh my god, here's all the things that I missed when that happened." Well, I love the unpopular opinion. So, marketing teams hate you, but people solving hard problems love you.
That's okay. I'll I'll take the customer liking me more than marketing. No offense to the marketing team. I'm in market. I know. And yeah, we got like the one to two minute uh remark. I want to hear like some fun stories like u probably half of the audience like newish to data bricks doesn't I know you talk about the founders experience but it really started small you were all like buddies and then like how did how did no pun intended the spark happen?
Yeah. How how did that come to fruition? Well, look, I think I I would love to tell you that there was binary there was this big bang period in time when it all happened, but um I think look in the early days what it would say is like even very early you could tell that you were tapping into a key need that people were having and the product was still lagging behind the vision. But realizing that those bets that we had made right which is bet on the public cloud, bet on open source and you know bet on um you know the fact that data and AI needed to be brought together in one world.

16:50 First customers

Uh I still remember you know the first uh alpha customer that we ever had was you know conviva working with them. I f I remember the first beta customers we brought in and I still remember you know uh fun fact that the first customer we ever signed was super excited was you know my fitness pal and it was like you know they they came back and now I guess part of Under Armour but it's like they came back and seeing them say this is where we think the world is going there isn't something out your guys' vision you just need to execute on it getting that reinforcement in the early days I think was the exciting part then it was like how do you just keep going and build around Okay, awesome.
And we have like less than one minute or I guess we're done. You're asking us to toss. So, thank you so much. Congratulations on the growth and all the uh the wonderful community that you've been building. Of course. Thanks for having me. Appreciate it. Appreciate it, guys. Thank you. Thanks so much. All right. So, now we're going to toss to ask DBS SQL. I want to make sure that everybody out there knows about datab bricks SQL.

18:26 Databricks SQL

It's now the fastest grown data warehouse in the whole world. More than just a traditional legacy data warehouse. It's a serverless intelligent data warehouse. So we at data bricks web zero customers a little over three years ago to now rich diverse group in every industry of over 10,000 customers. and they moved there uh because of its incredible price performance as well as the incorporation of AI all governed in the unity catalog framework for the best uh security.
So for performance, we typically see four time performance even at a lower price. For serverless, this is one I personally have experienced where you can get near instant and elastic compute. So when I give demos or I'm working on my own projects, I don't have a data warehouse started, it comes up in just a couple of seconds, which years ago would have seemed nearly impossible. And it also has the latest and greatest everything AI or large language models all built into SQL.
So for example, one of my favorites is AI functions where in the SQL command you can do things like text translations or sentiment analysis of every record that you're scanning through all simply within uh SQL and the databicks uh SQL. So if you have a legacy data warehouse, we have a multitude of partners. We have in-house migration tools that will help you along this process. So, thanks

Learn more about the Databricks Data and AI platform.

The information provided herein is for general informational purposes only and may not reflect the most current product capabilities or configurations.