The following interview is a conversation we had with Tommy Dang Co-Founder & CEO, on our podcast Category Visionaries. You can view the full episode here: Tommy Dang, CEO of Mage: $6.3 Million Raised to Build a Modern Replacement For Airflow

Brett

Hey, everyone, and thanks for listening. Today I’m speaking with Tommy Dang, CEO of Mage, an open source data pipeline tool that’s raised over 6 million in funding. Tommy, thanks for chatting with me today.  

Tommy Dang
Thank you so much for having me. I’m super excited for this. 

 

 

Brett
Yeah, no problem. So before we begin talking about what you’re building, let’s start with a quick summary of who you are and a bit more about your background. 

Tommy Dang
Yes, would love too. So a little bit about me. I’m a Bay Area native. Grew up here, went to school here, went to school in Berkeley nearby, and then I lived in San Diego for a little bit. Started a company with a friend down there. Eventually I made my way to joining Airbnb early on in 2015 there, I joined as a software engineer, worked on dev tools, data tools, data infrastructure tools like Airflow. Worked there for over five and a half years. Even got to work really closely with our CEO, Brian. Learned a lot there. And then in late 2020, in December 2020, I left and I started Mage. And we raised our seed round early last year. In 2021, built a team out, launched a product, went to general availability. Then we decided to eventually open source one of our core technologies that we built. 

Tommy Dang
And that’s what you see on our GitHub right now. So we open sourced that earlier this year, and it’s been catching on like wildfire. 

Brett
Very cool. And to zoom in there a little bit on your time at Airbnb, what was that like working there in 2015 and how did you see it change over that five year period? 

Tommy Dang
Yes, great question. It was really awesome. Back early in the days when I joined. This might sound big to you, but it’s actually quite small. There’s less than 200 engineers and the company was less than 2000 by today’s standard. That sounds like a giant company, but it wasn’t. We were only on a few floors in the building and everyone knew each other. It was really fun. One of the core values of Airbnb was be a cereal entrepreneur. And when they mean cereal, they mean the hominym for the food, cereal, but also cereal serial. And one of the core values, how that lived out is you just do anything you need to do. You treat the company as an owner, so you take it upon yourself to do what you got to do to hit a goals, to help users to help customers, et cetera. So there’s this really high level energy aura around the company of just being really scrappy and it was really fun. 

Tommy Dang
You got to do just literally anything you want. You can move to any team, work on any project as long as it helps achieve our company’s mission. And that was around for quite some time and it really gave a lot of flexibility in what you can work on. And that’s why I was able to work on I started working on a lot of data intensive tooling and products just because there was a huge gap in the company at that point. And so over time, though, of course, every company grows. By the time I left, we had about 2000 people in the engineering and then the company was about 8000 people globally. So people start getting more specialized, teams start getting more specialized, there becomes more processes. So things eventually slowed down. And you were on a team for a very long time. You were working on this one thing for a very long time, but that was a natural progression of just any large company. 

Tommy Dang
But man, in the early days it was really fun. 

Brett
That’s cool. And being a software engineer in the Bay Area, working for Airbnb, I’m sure you were well compensated for that. So it must have been a bit scary to give that up, I’m guessing, and to go and build your own company. So what was going through your mind when you decided to make that leap? 

Tommy Dang
Great question. Just anything scary going from something stable to not stable, going from a known quantity to an unknown quantity? Or just familiarity and being somewhere for five years. You get in the habit of things, the cadence of things. But one thing I remind myself is I always ask myself, am I doing the most exciting thing possible? Whatever I’m doing right now, is it consuming 100% of my mind share? Is it consuming all of the energy I have? And when something doesn’t, that’s when I tell myself what else is what is pulling that other 10% over time? What is that? Maybe it’s a hobby, maybe it’s another idea, maybe it’s a passion, like volunteering. It can be anything. So then over time, you start having to ask yourself if that area of my life or somewhere else out of my life is pulling more and more percentage of my mind share and energy. 

Tommy Dang
Should I start looking into that? Should I start adding more time? Should I start actually shifting my entire energy and focus to that? And so that’s what started happening. And just working at RV, I was able to witness a lot of pain points and a lot of pros and cons of lots of different tools. And I got to work with lots of engineers and developers and I saw a huge opportunity for what we’re building. And so that started taking up a lot of my mind share and where. 

Brett
Did you learn that mental framework and those types of questions to ask yourself? Because it’s really fascinating. I can see how that’s super useful, but where did that come from? Did someone teach you that? Did you read it in a book? Did you just make it up? What’s the origin story? 

Tommy Dang
I’ve never had someone ask that question. That’s a great question. I’m thinking back. So at my first startup, we started in San Diego with a friend of mine and it was a two sided marketplace for student housing and that consumed all my mindshare at that time. So I didn’t learn it there because then eventually we shut down and I joined Airbnb. I think I developed this mindset at Airbnb because when you joined, I joined this early team called Business Travel and it was a really fun team and I was on there for almost a year. And towards the end of that team, my mind started wandering a little bit in the sense that I shipped everything, all the features, all the stories, all the epics that I could faster than I can even receive them. So then I started thinking about, okay, how else can I contribute to the company? 

Tommy Dang
So I started participating in different forums, I started helping out people on other teams. So my mind started wandering there. And then I was recruited to join this team that Brian Jessica was putting together called Magical Trips. And then that consumed 150%, 200% of my mind, right? So all my energy was poured into that and that actually was able to extract almost my entire being, my entire energy for quite some time. And then after that was in a really good place and there’s so much, many more people involved in that business because we launched it as a new business. That’s when I started thinking about, okay, where are some other ways I can pour my energy into? And that’s when I started some dev tools and worked on some data tools at Airbnb as well. So I think I developed that over those five years because you got to constantly check where your energy spent and revitalize yourself if you’re in this or in some organization, that’s how you have a long, prosperous career in a large organization. 

Tommy Dang
But then towards the end of all of that, I started realizing, you know what, there’s something else that’s been on my mind and I’ve been dedicating more energy to it. And that’s the idea of me. 

Brett
Very cool. And two other quick questions you like to ask just to better understand what makes you tick as a founder. So besides Brian Chesky, is there a specific founder that you really admire the most and what have you learned from that founder? 

Tommy Dang
He said besides he was that one that I was going to talk about. Okay, I will take another airbnb ex airbnb person. His name is Joe Bot. His actually name is Joe, but people call him Joe Bot because there was another Joe that was a founder at Airbnb, and he was a really early employee at Airbnb, and he grew to become a pivotal cultural leader and a leader there. He’s one of my close mentors, and he’s also an investor in US. But he taught me many lessons. But there’s two that I repeat more than others, and I repeat this to myself, I repeat this to everybody I meet. One. The first one is don’t define your responsibility. And we’re hired on for roles. Roles are very, I would say structured it’s. You are this role. You are an engineer, so you only code or you are a marketer, so you only do marketing. 

Tommy Dang
Well, this breaks that paradigm. You don’t let that role define your responsibilities because your responsibilities at the end of the day, are to help users, help the customers grow revenue, help others at the company to then help them achieve their goals. So that’s not ever defined in a role. Yeah, you join a role and it has a set of, let’s say, five responsibilities, but you should go beyond those. You can find ten other responsibilities, 30 other responsibilities. And that really revolutionized the way you think about yourself and others. And then the second thing he taught me was do things you’re unqualified for. And that means, yes, people, you might be brought onto a job because you have these certain qualifications, but you know what? You can actually do jobs that you’re not even qualified for. If you think about a lot of the jobs we bring on, you might not have all 100% qualifications, but you know what? 

Tommy Dang
You can still do that. You grow and expand into the space that you need to. And in addition, they say if you wait until you’re ready, you’re going to be waiting for the rest of your life. So no one’s ever qualified to do many of the great things that they end up accomplishing in life. So don’t let that hold you back. And that really helped revolutionize the way I think about things as well. 

Brett
Yeah, those are such amazing lessons. And what about books? It has to be a non Airbnb related book since you’re going down that path. So non airbnb related. What book would you say has had the greatest impact on you as a founder? And this can be a business book or a personal book. 

Tommy Dang
Okay, so there’s a lot of books out there. Okay. But one that’s top of mind for me that I really love is called Team of Teams. It’s by General McCrystal. He was the head of the Joint Special Operations Command in the Middle East. And basically it’s called Team of Teams because it’s Joint Special Operations Command. That means there’s special operations from different parts of the military, and they’re all disparate. But in order for them to be able to work together to the maximum capacity they need to be very interoperable with each other. So the book talks about how do you get a bunch of teams disparate teams that have nothing to do with each other, to be moving on a cadence of as if they were one team. So we think of a team as individual people. Okay, that’s great. I think we got that down. 

Tommy Dang
But then how do you compose a team of many different teams? And so, at a very high level, I think that was really awesome for me because it helped me think about incentives, goals amongst individuals, and again, teams or entities. And I think I read this when I was at ARMB because were building tools for other teams at every B. And once you get to a sufficient size, you have to start thinking about what makes the other team tick, what are their goals, what are they trying to achieve? How do we operate in a way that helps them win? And that was basically the summary of the book. You basically help the other side win, the other team win. And that’s how you create a team of teams and everybody’s helping each other win. I’m helping you win, you’re helping these other three people win, and these other three teams are helping this one team win, and so on. 

Tommy Dang
And so that’s how you create this scalable incentive model where all the teams operate within some other team’s best interest. 

Brett
Very cool. It’s always exciting to hear a book that I haven’t heard of before. I would say 90% of people come on and just say the hard thing about hard things. So I like hearing new books, and I’ll pick up a copy and read it over the weekend. Thanks for sharing. 

Tommy Dang
Thank you. Yeah, I’ve also read The Hard Deal Hard Things. That’s a good one too. For sure. 

Brett
Nice, man. Let’s switch gears here and let’s dive into Mage and what you’re building. So to start with, can you just take us through the origin story? 

Tommy Dang
I know you touched on that a. 

Brett
Little bit in the beginning, but let’s go deeper into the origin story and then talk about what you do at a high level. 

Tommy Dang
Yes, absolutely. This is actually a great story. So at Airbnb working on lots of data tools, initially, I saw an opportunity we had disparate tools for pulling data, extracting data, transforming data, storing the data, building training sets, reusable data sets, feature engineering, also deploying models, training machine learning models, et cetera. And a lot of these tools I saw were built for really specialized roles like data scientists, data engineering, things like that. And so I had the idea that, okay, what if we took some of these tools and made it more accessible to a wider technical audience? And I don’t mean making it accessible to anyone who can’t code, but there’s a lot of people who can code out there, but they just maybe didn’t go to school for machine learning or didn’t spend all their years in data engineering, et cetera. So we left and I built out this end to end machine learning platform that composed of a lot of these data tools. 

Tommy Dang
Think of a data pipeline tool, think of a feature store, think of an online model serving tool, AutoML, things like that. And so we built this Cloud hosted and we launched in early 2022. And it was great. We have paying customers, we had thousands of users sign up. It was great. However, were targeting early stage companies, series, ABC, Seed Stage, et cetera. And what we saw was, although there’s an appetite for machine learning, a lot of these companies, they want to use machine learning, everybody wants to use machine learning. But what we found is they actually struggled with a more urgent data challenge early in their journey. And it is just the movement of data, the transformation of data, the integration of data. And so we took one of the core technologies we built internally, because our platform behind the scenes was composed of several pieces. 

Tommy Dang
We took the data pipeline piece, spent a couple of months working on open sourcing that and we open sourced it because we just had the hypothesis that hey, people need this earlier in their company’s journey. And we validated the hypothesis. And a lot of companies started picking this up and using it in production way more than our Cloud hosted version of our Cloud hosted ML tool. So that’s when we started focusing a lot more on this open source tooling. 

Brett
Very cool. And talk to me about traction. What type of traction have you seen so far? 

Tommy Dang
Yeah, absolutely. So after launching back in June, we started seeing more stars, we started seeing more contributions, more issues, open people joining our Slack companies using in production. So we have over 2000 stars, over close to four to 500 Slack members in the community. We have over a dozen or so companies that we know of that are using it in production. There’s quite a bit of others using in production as well. Just folks that are mentioning it here and there. But we don’t track telemetry data at the moment. So there should be quite a bit more usage out there that we just aren’t even aware of. 

Brett
And how do you think about market categories? So is this a data pipeline platform or what’s the actual category? 

Tommy Dang
So we position ourselves against Airflow. So we are a modern replacement for Airflow. I think that’s just the easiest way for data engineers to understand what we do. However, Airflow is in the category of an orchestration tool. Now, orchestration tools, typically all they do is think of them as a conductor symphony. They just say hey, you play this, play this. But they don’t actually play anything. And so that’s what those orchestration tools do. Like Airflow. For us, we are a data pipeline tool. So we build data pipelines. What are data pipelines? They are a series of steps in some sequential order. It doesn’t have to be exactly one for one. And they move data and then they mutate or transform that data along the way through that pipeline. And then that data ends up somewhere else. So we basically focus only on data workflows or data use cases. 

Tommy Dang
Airflow can literally orchestrate anything because you build Dags and Airflow to call this API, to do this, to maybe sometimes do something with data, but you’re always calling out to different APIs that can do literally anything. We’re just focused on data. So yes, we call ourselves orchestration just for simplicity’s sake. But we are much more specific than that. We say we are a data pipeline tool for transforming and integrating data. And because of that focus, we can do a lot more. There’s things like we could do automatic versioning partitioning, backfilling. Every step in your pipeline produces data. We can have really specific monitoring around data, things that are not first class citizens in a tool like Airflow. 

Brett
And if you’re going head to head against Airflow, obviously they’re a big established company. Why do you think you win those deals? If you are going head to head, what are those points of your product or features of the product that are making buyers and users want to use you instead of them? 

Tommy Dang
Good question. There are four of them. So after working on me for a long time, getting a front row seat into seeing how developers are using Airflow, I’ve seen, I was there when it was open source. It saw it go from ten Dags, tens of thousands of Dags. So we take some of the best parts and then we revolutionize all the worst parts. So there’s four main key differences. One easy developer experience. As many people that use Airflow, I haven’t met one person that said they love working in Airflow. There’s a huge opportunity to improve that developer experience through great UI, through better design principles, through better coding design, things like that. The second way we also differentiate ourselves is engineering best practices built in. And how that plays out is we actually enforce more of a modular pipeline. As you build out your data pipeline, every step in your pipeline is an individual file. 

Tommy Dang
And that makes it really modular, reusable, easy testable, and you can even write inline data validation. The third way is we treat data as a first class citizen, I explained that earlier where all we do is data, movement of data, transformation of data. So we have a lot of these built in, functionalities like versioning partitioning backfilling, things like that. And then the fourth way is scaling made simple, airflow can scale up to tens of thousands, nearly infinite dads and tasks. But it also takes quite a bit of dedicated resources to do this, to maintain it, to upkeep it, to debug it, to monitor it, et cetera. Everybody had quite a few data engineers to do that, but not everybody has a team of 20 or 30. And so we make it really simple for even one data engineer to manage thousands of pipelines. We also make it really easy to scale up the processing of the data. 

Tommy Dang
So you can actually run your data transformations or processes on the compute that’s running mage inside dockering containers, or you can leverage Spark, we have made integrations with Spark. Or you can write SQL, mix and match SQL in your pipeline alongside Python, and that SQL will get executed in your data warehouse leveraging the compute there. So we make it really easy to scale up. 

Brett
And do you think data pipeline is going to become an established category that eventually becomes full of other companies as well? Or how do you think about categories in general? 

Tommy Dang
So, data pipeline is actually a very broad term because anybody who does anything related to data in some sort of workflow will say the word data pipeline. They mention it, but they don’t orient themselves all around it. And so we do that just because everyone I feel like a lot of the tools out there want to broaden their applications. So let’s say there are tools out there that do general orchestration or generic pipelines. You can expand that horizontally to do machine learning, for example, or different types of pipelines. We didn’t want to split our focus in that sense, so we are only focusing on data. And so you can see that the current category is there’s also data integration. There are some that only focus on data integration, but we feel like that’s too much of splintering and then being too precise. So there’s this whole idea of unbundling Airflow or unbundling some giant monolithic tool. 

Tommy Dang
Well, the result of that is you have hundreds of these other tools that split up and do one thing only. Now, I would argue, yes, sometimes that makes sense, but in some cases, some things do not need to be unbundled. Some things work better together, something’s synergized together. And for us it’s actually the data integration data pipeline piece because right now in the market you have these split, you have some that only do data integration, so move data from one place to another. But some that just do transformations, do process, reprocess, recalculate, aggregate data that’s already in a certain destination. And so we actually see huge synergies between the two. So we started combining the two and we’ve seen lots of great use cases out of that. So in terms of category, we’re going to see some consolidation of tools. You’re going to see IBC data integration tools combined with the data orchestration tools. 

Tommy Dang
You can also see other data tools in the data stack combining as well. So I see some consolidation in this, and we might have a new category out of this, but for now we play around in the orchestration category, makes. 

Brett
A lot of sense. And in terms of getting in front of developers, getting them to use the product, obviously there’s been a huge amount of funding over the last few years. Go to DevTools. So what are you doing to stand out and how are you going about marketing to developers when developers are very allergic to marketing? 

Tommy Dang
Typically, exactly. I know what you mean with that. First off, how do we stand out? We believe strongly in great design. And what I mean by design, I don’t mean how things look. Design is how things work and how things feel. And this can come from coding. How do you feel when you write code this way? Or how do you feel when you have to click a bunch of buttons to do something very simple so it permeates throughout the code, throughout the product, even off the product. How do you feel when after you launch a data pipeline and you have to go to sleep and walk or go to lunch? Do you feel secure? Do you feel like you know that if something goes down, you’ll be notified? So how do you feel about this? Design is super important to us and one of our core design principles. 

Tommy Dang
Actually our number one core design principles is easy developer experience. And this is our sole focus and this is how we differentiate ourselves. And that’s why you see a very rich user interface. We ship with the user interface and that’s our primary focus. Now you might think, is this low code? No, code. It’s actually very high code. When you write the steps in your pipeline, you’re writing a lot of Python code, but everywhere else we augment you with a nice intuitive user interface, which I don’t see many others really focus on. And so we really lead with that. So that’s how we stand out. Now in terms of marketing, selling to developers, I’m a developer myself, so I know exactly what I don’t like. Right? I don’t like being sold to all these things. It’s just as a developer, we like to read docs, we like to try things out on our own. 

Tommy Dang
We like to come to the realization like, wow, this is magical ourselves. So what do we do? We do a lot of things that don’t scale at our stage. We do things that don’t scale. We simply talk to everybody. We meet with everybody, we go on zoom calls, we meet in person, we go to in person meetups. We know everybody that we come in contact to that uses Mage, where we’re not afraid to hop on calls to help people debug. We’re not afraid to pair program with people, we’re not afraid to work on issues that they have on the spot right away. So we’re definitely not afraid of that. And we believe in one one relationships. And that’s how we break through the noise and reach our end users and. 

Brett
That makes sense for the end users. What about the actual buyer? So who’s making the actual buying decision? 

Tommy Dang
Yes. So we are open source and we don’t have a cloud hosted version right now we’re not focused on our Monetization strategy which we’ll start executing our Monetization strategy in 2023, which would include a cloud hosted version, hybrid deployment where they deploy in their cloud and we charge management fee and then enterprise specific features. So right now it is all bottomed up. The developers, the data engineers are using it, trying it out, loving it, bringing it in and getting their team, their company to adopt it. So for them they are currently the buyers of the tool because well, they pay zero. They pay in time and effort in setting up. But that’s the cost that they have to pay themselves and so they’re willing to pay that. So that’s how we are approaching that now. How will we eventually have folks who want to use our paid solution, et cetera? 

Tommy Dang
Well, we’ll have internal champions. The data engineers, the community that we build really love this open source project and it really helps them, helps make their life and job easier. So they would be internal champions when it comes time to having to speak to let’s say engineering manager or VP of engineering and sharing with them. Hey, you can use Mage as a cloud hosted version or the enterprise features. Well half of your data engineers are already using it or have used it in past companies or in the past. That makes it a lot easier for that decision maker to make a buying purchase. 

Brett
And are you nervous about turning on Monetization? I think Figmo is the most recent example I heard of where they waited years until they did that. I don’t know if that’s accurate, but that’s something I had read online was I think they waited like three or four years getting the community to use it and then they turned on Monetization. So for you, does that scare you. 

Tommy Dang
To take that step? 

Brett
And what metrics are you evaluating to make sure the timing is right? Because I’m sure timing is everything there. 

 

Tommy Dang
Great question. We aren’t afraid because we’ll always have a free. The open source project will always be free and we dedicate time and energy to make so. Part of easier developer experience and scaling Made Simple is a single developer being able to manage it themselves, deploy themselves, manage themselves and not have to pull their hair out and stay up all day to maintain it. So we make it really easy. So we aren’t afraid to turn on Monetization because we are not going to leave those folks who are self hosting in the dust. Those are our number one focus to continue build community. But to turn on Monetization, someone who’s already using it doesn’t have to be afraid that oh now they’re just going to mage. Is just going to focus all on making money? No, we’re still supporting those that are running it themselves. And the monetization part, the cloud hosted version is for those that might even not have Indian data engineers. 

Tommy Dang
Maybe they just want to use something that’s already there. Maybe they don’t have a cloud host, AWS, GCP, Azure, et cetera. And so they just want to get started or play around with it. And also monetization off of enterprise specific features. That’s even more so for large companies who really need specific features where we don’t believe majority of our community needs them. And so for some metrics to signal when is the right time, well, we’ve been getting pulled from the market asking, hey, do you have a Cod hosted version that I can just use without setting up? I don’t have any data engineers, for example. So we’re getting more and more of that pool and it’s just a matter of us executing on that, taking the time and prioritizing executing on that. 

Brett
That makes a lot of sense. And in terms of your go to market so far, what would you say has been the single greatest challenge that you’ve had to overcome and how do you overcome it? 

Tommy Dang
Yes, it’s trust. This type of tool isn’t some task management tool where you can just try it out and then move on. This is a key critical component in your data architecture, in your infrastructure, and it deals with your data. It moves it around, it transforms it replicates it. So it’s super important to be scalable, be reliable, have lots of monitoring in place, and do exactly what it promises to do. So how do you overcome the trust? We’ve only been around for two years and something like Airflow has been around for seven, eight. So overcoming that trust is key for something like this. So how do we overcome it? It goes back to how we break the noise is the one one relationships. People know the team, everyone who’s using it knows us. They know the community, they know the dedication, they know the pedigree, they know the history, the experience of the founders and of the founding team. 

Tommy Dang
And so that really helps bridge that trust gap. And it’s a snowball effect. The more trust that we build amongst our early community, the better use cases they deploy, the larger scale that they deploy that then feeds into case studies, stories, testimonials for the next group of folks, the next cohort of companies and teams that might be hesitant. They hear from that, they experience it themselves, and then those become champions and those become diehard fans and so on. 

 

 

Brett
Makes a lot of sense. Now, last question here for you. If we zoom out into the future, what’s the three year vision for Mage? What’s it look like three years from today? 

Tommy Dang
So three years from today, we see a world where Mage is the go to data tool for early stage companies, mid sized companies. It’s a tool that you think about and you spin up as soon as you start a company, as soon as you have any database set up. We help these companies build best data engineering practices from the get go. Also, we see ourselves being the end to end data pipeline development in the cloud. We support running locally and also running in the cloud. But what we have seen, we recommend to companies to use Mage end to end in the cloud. You can develop in the cloud, but also you can run it in production in the cloud. And this actually helps with developer productivity and developer velocity because you get to build out your pipeline on cloud resources so that then when you deploy it in the same cloud environment, but just in a production environment, the errors, you’re not as surprised. 

Tommy Dang
And so you get a lot of parity between the two environments. So we really see a world shifting to that and we have lots of support for that right now. Another thing that we see ourselves is we love doing being the dirty and boring plumbing behind the scenes for companies. We want to get to a place where everything is so easy, so smooth and so transparent that you even forget that we’re here. We’re just in the behind the scenes doing all the plumbing for you and this just runs very smoothly. There’s no complaints when things break down. That’s when you start noticing something. We want to almost people forget about us because nothing breaks down. And your data pipelines are just so well secured. You have data validation, data quality, data contracts in place where you set it up once and it just runs for eternity. 

Brett
Amazing. I love it. Tommy unfortunately, that’s all we’re going to have time to cover for today’s interview before we wrap. If people want to follow along with your journey as you continue to build, where’s the best place for them to go? 

Tommy Dang
Yes, please, everyone, go to Mage AI chat and join our chat. Join our online Slack community. All of us are there. We’re very responsive. We’d love chatting with you. You can DM me on there as well and I’m happy to hop on a call and we can just get to know each other. Awesome. 

Tommy, thank you so much for taking the time to chat. This has been a blast, getting to know about what you’re building and just getting a better understanding of who you are as a founder. So thanks again and we wish you the best of luck in executing on your vision here. 

Tommy Dang
Thank you so much, Brett. I hope I can come back again. See you. Keep in touch. Bye.