About the talk
2020 has been one of the toughest years for fintech companies and the financial sector in recent memories. However, the adoption of AI and data technologies haven't slowed. This talk explores the impacts brought on by Covid-19 and how the they are driving game-changing AI and data trends across the fintech landscape for the coming year and beyond.
About Data Application Lab:
00:00 02:08 2020 Recap
00:00 04:35 Customer Behavior
00:00 06:16 Brick and Mortal Businesses
00:00 07:10 Surge in Data Volume and Availability
00:00 08:49 Trends in Data Tools
00:00 11:33 Trends in AI + ML Application
00:00 16:10 Explainable AL
00:00 17:16 Financial Inclusion
00:00 19:33 What in Luna
00:00 20:41 LUNA - Al-Driven Business intelligence to Mew Mainstream Latinx Economy
00:05 Introduction with Cho-nan Tsai
Okay, so let me get started on the topic today is top a I ended up strands in feng Tech rifle 2021 and this is a very interesting topic that I've been thinking about for a while, right? And I love to get deep into it. And I think obviously, you know, I'm going to talk more about fintech because that's, that's the area that I working, but I think a lot of it actually applicable across different areas, so happy to do the digging more during the Q&A session, but let's go through it quickly. I want to share a little bit about the company that I work for
and my, my personal background. So, I work for Community Financial. It's basically a next-generation platform and we specifically focus on providing loans, and Lending Solutions, and financial services to the Latino space in the US. Okay, I've been Working in technology for over 20 years, my background has always been intact. And in computer is sort of my thing. Early in my career I worked in Fortune. 500 companies, look for a lot of like, you know, big name companies and manages the ton of multimillion-dollar systems
more recently. And my career I've been sort of working, in startups, know, I help fundraise manage teams you know mostly in and near capacity. But obviously for this company right now, you know, I'm very very much in the trenches with the data as well and and so you know my experience that is kind of you know spread across the entire data digital media attack and security as well. I'm a graduate of UCLA and I went to Columbia University for MoneyGram So that's kind of my background
International. Before I start kind of going through the trans, I like to take a moment just to allow us to sink in and think about what happened this year. All right? So 2020 has been a very, very tough year for not just from sex but I think in lot of a lot of business sectors in general, right? And and we know, you know, hands down like coronavirus was deleting problem of of it right back in February. You know, this coronavirus took off in the US we
had the first coronavirus related deaths reported in February, right? Having this debate, you know, should we stay open or should we actually closed office? Go office and work from home, right? It was a lot of debate and going back and forth and before you know it, you know, or employees was like, coming up with excuses to work from home. The one of our executive members of the team said, hey, you know, we we better monitor this thing closely because our reputation and credibility is at stake, you know, we don't want
to fall on the wrong side of the history here. So we then right there and there we set it to, you know, move to remote working and the next day. And and then and that was kind of like around March this year. And then one by one, you know, the stack the states went into lockdown and of course we're glad that, you know, we had to pee pee and then go to work from home. And so But things could see that happen, right? Black lives matters took place in the summer, that was a lot of looting and then people protesting, right? And then this year, we
had the worst natural disasters in the US join fires that we had was the largest ever in California and there's all kinds of things going on, right? And then, on top of that, of course you have a political climate that is very, very turbulent as well, right? With the elections, coming up in a few weeks away and, you know, I want to charge every call The Boat if you haven't, please do so. But anyways, let's move on. So what does Aldi's in o mean at the end of day? I think from Athena perspective, we can think about it. As you know the truth.
Human behavior is actually changing rack radically. And how radical is it? Let me illustrate with the story. So, I have a neighbor is name's Max is like a second grader kid, you know, he would come over and play soccer with us, and we were still kind of playing soccer a little bit, even after covid happen with the masks and social distancing and everything, and then after awhile, you know, the kids start stop showing up at my house and I was like, Curious, what happened? So I went to knock on this door and it was nobody there, right? So I started
asking everybody in my neighborhood. Okay, what happened to that family? They're not there. And then one of my neighbor came and told me, oh, Max, you know what? He's too cool for school. She's parents and I took him to move to Cary be. All right, so that's how radical things are. Like, people are either moving away from the life. They were leading before covid, or they're just going to stay at home and not going anywhere, right? So people are just plain different type of things member of back in March. April people, hoarding toilet papers, right? People love cleaning products and
they still do right? And maybe think about it, you know, how Ubly you're buying stuff online or your groceries, you know, of making just one stop at Costco. Instead of making four five different stops to limit your exposure. So, totally change, right? And this is on the personal side. What about on the business side? A lot of physical stores near not able to survive. So this is the day that I took from Yelp. So more than 41% of shops in the US have Community clothes, I think the number is probably higher. Now, you know, cuz right now is October and so there's a ton of
changes, right? That's going on. But at the end of day, what does that mean for theater scientists and did a professional? Well, if you ever worked in one of these companies, write that focuses on building models, right? Trying to predict certain behaviors, you can take that model and put it in the trash can for now. Because a lot of things for a lot of things to right now I'll totally different than not the same as they used to be, right? So we need to have a better way of going about it, you know, because they are all these different things
coming on board. What are the trends that I've seen so far this year there? I think they will continue to to go on in the coming year as well as, you know, there is a rise in data balling. So right now, just to let Estates latest statistics that I have everybody is generating about to megabytes of data for second, right. By the end of this talk you're going to see gigabytes of data and because this is streaming video. So maybe penso's you know, tens or hundreds of gigabytes and told her I could person and And, and you have a lot
of this day. Going on and they are now being made available as well. Write Us by companies of business, because they need to collaborate right with each other. So you can see a lot of these API enablers. They're out there right at 9. Business, too kind of work together passing on information to one another. Do you know to actually make more profit or to actually improve certain conversion rate of certain funnel, right? And this is just a digital stuff. What about the analog data? While you're there, are some vendors out there that specialize
in taking analog data at stuff is on paper or maybe Insidious be converted that into Chase on so that you can feed that back into your data models. So overall, there is rice of data volume, availability, you know all over the place, right? What kind of the first train and with that, you know what, you going to see you with the rice that you going to need more tools right to to manage all that. So I've kind of categorize the food in three packets of data security. In
adjust. Now, the previous speaker, Samuel Burger in touch on this topic in that and I I totally agree. I think there is definitely going to be a need for this. You know, why? The steps they have is there's be more cyber attacks in the fast first half of 2020 than 2019. Okay. And that's stressing that the number that we're saying. So what are you going to do? You know, you're going to you know, see some of these tools, they're interesting, right. And I think Samuel earlier alluded to some of these tools are great and I think a lot of them are
good for fighting or preventing against external to text, but what about in China? Text. What if you have employee to enroll cry? Because you couldn't give him a race to the covid-19 or whatever, right? So you have been just like Doc trays that, you know, Bill's Rai model that monitors all the user behavior is on your network. So whenever there's any of novelty right to flax will go off, right? Does that mean we'll get alerted and then you can start shutting down accounts. So that's what that way on the rise. Right. And then lollies tools, right? I have some of that logos here.
You know, SackMan snowflake which is a cloud-based data warehouse while they're just going to the cloud all together in the past. You may have seen companies big and small old going to data center is right, building out their data warehouse, you know, hosting their servers. That's all in the past. Like you wanted to move quickly and Changing Times, you know, you got to be called base and everything has to be moving at light speed. Write the second offense has just recently got caught acquired by twilio for about close to 3 billion dollars, that's what the be, right? And
that's kind of money for a company that focuses just on customer data platform, right? And last but not least, you know, you got the other privacy issues that everybody is concerned about you. Going to always say that everywhere, right? But where are the tools that will protect consumers, and that helps right businesses by fintech companies that manage all this data, you know, accordingly to the law, right? So, you will see your eyes of these vendors are going to play a big role in the ecosystem. so, those are the tools now. Let's talk about
applications a little bit. I want to start off with the story of mine. So I remember back in the day, when I was a kid, you know, my mom took me to this poster Bank, you know? And I remember like the, you know, the moment that I went in and brought my piggy bank, you know, I can deposit my money in the bank account and the teller in a writing that amount on a little past book with the signature in the stamp. Well, that was back in the 80s that mean annual today. They ain't got time for that, you know. They want beautiful stream my experience almost
Amazon like, right? So you go in, you have all this predictive personalization you not to your age to your profile. So the stuff that comes up on the products that you would like right away, right? Doesn't matter whose Financial products in Mishawaka, whatever it is. It's all heavily personalized to your needs, right? And the stuff like a co-writer voice, communication tools are built on top of a, i n. Modules, you know, right now you can use it to buy stuff on Amazon and tell you to turn your lights on and off. But there's nothing stopping it from
making Financial transactions. You can say, hey, you know, go and move my money from my savings to checking's and pay my bills to The Gardener, right? So there is all these changes are coming, which I think it's exciting and futuristic, right? So that's the first of these applications that I thought would be interesting. And the second is Frosty text, right? With so many, you know, phishing scams out there Frost, you know, this year, we had obviously a lot of people trying to, you know, defraud PVP and and all
these other initiatives as helping businesses, you know, and even our own company. Like we seen you know a lot of Sprott applications that were coming in. So the most effective way to shoot through all the transaction right is obviously you're with machine learning model, right? And Nnn with Decent Behavior monitoring, you know, you can raise the flag sooner, you know, brother and later when the damage is done, right? And the third thing is modeling, right? I think, even though it's tough, don't know that we have
built before the models. Don't, I'm not as relevant anymore. You're supposed to be building new models to write to accommodate The New Normal, right? So I think there's going to be strong applications continuing, you know, in a credit and insurance text space, right. You may start seeing like these very personalized credit terms or Insurance terms is a tailored to your lifestyle, right? You need to go low-risk, you know, you pay lower premium, right? And things like all my customers chair and that's always been there before. I think there's going to be more off the shelves or
solution that you can tap into and get a sense of okay, what the trans going to look like? His dad has all available right now. What does that mean? Write a lot of things tech companies or companies in general right now. They got to be more lean and mean. You know why? Because businesses You know, is The New Normal, right? You consumers are not behaving, you know, normal anymore. So you, if you can improve your conversion rate, this by 1% and with the right machine learning model, AI, Then maybe you can break even or
even make some profit, right? There's this great application here. And then that collection, of course, you know, with millions of Americans that has lost their job. You know, you're going to see people in a behind on their payments, right? So in the past, of course, you have people on the phone talking to the people that are behind but that's a very slow and inefficient way, right? Of, you know, getting that, you know, probably pay. And so the good thing is, you know, the machine learning and AI tools out, there is mature enough to start doing things like validating,
you know, people's promise to pay if they're really mean it, you know? So that's one. So the trend that I thought would be would be very important to cover. Here's well and nothing is, you know, this expandable AI for credit approval. I think you know people have been talking about it for a while. I think you're going to be seeing more of this. Especially in the sting techspace, you know in the past. Yes giving credit you can go through the roof based method which, you know, I think it's okay during normal times. But in sometimes, you got to be very
creative, you know, with your models, right? You need to be able to get that extra Edge so that you can prove the conversion rate and so you can give an answer to a credit application but don't just give an output. You know, you have to give reason, right? Hey, I didn't I this case because you know, he's Revenue was below a certain threshold, right? So you'll be able to build some AI model that can actually do that, you know, very succinctly so that you are you know, the cooperating your business according to the legal boundaries of a new off the business.
So those are the main trains. Here's one more that I thought would be interesting. I think Jason said of alluded to this a little bit. I think he say something like, you know, there are these news players. They are not able to survive. It's true. If your Niche is very small, your business is probably your business. Probably is not going to buy Camino Financial has been very smart about this. We focus on the Latino space in the United States. There is Alarm and Bank individuals in the world. Okay.
One point seven billion for around the world. And in the US, the Latinos together, they generate about 2.3 trillion dollars, in GDP trillion with the tea, is huge amount right in the US were in the country by itself is going to be the 8th largest in the world. So this is a big Niche, a giant, right? And in the past, people say, hey data for the space is not available because Nana bankable there's no paper trail, no digital footprint, whatever? Right. But that's not true anymore right there. For this group is now becoming available in the last five, ten years out of data vendors, you know,
happened working to scoop up data from different places and even more. So you can build, you know, Approximation model to some of these things when you're using a surrogate in a models write, thank you can predict maybe the expenses or incoming or what-have-you right to make him act more accurate, even for unbank individuals. So this is another very interesting that I thought I would share. Now you may ask okay thanks Jonah for sharing. All these wonderful tips and trans.
How do you put it together? You know, there's just so many things that you can focus on and expand your business, right? So Thankfully in 2019, we started thinking about this problem. We wanted a framework that can help us, you know, scale-out business. Even during tough times a fast Changing Times. So we gave the name of this framework loona. So what is Luna anyways? Right? What Luna is also the name of this ancient pyramid, right? That's in Mexico. It's called the Pyramid of the Moon in English
and there's something very, very mystical about this parent to today. Nobody, nobody has any idea. About how the pyramid was built, okay? Given the technology they had, you know, in those days right? And they are people cycling experiences. When they climb this pyramid is closer to get to the top more spiritual energy. They feel right now is building a proprietary you know text that can be more insightful, right? And definitely more proprietary indefensible when you get closer to the top, right? And we have samurais everything here that I love to share with
you. This is probably the most critical slide of the entire presentation but is very applicable. Write two different sex. The older we don't, we wrote this with fintak latino economy in my butt and Nina de luna is the AI driven business intelligence pyramid at the very bottom Right? That is the foundation of everything. Know you have cloud-based infrastructure, third-party, software the best-in-class. Don't build it from scratch. You buy for that. You pay for that for instance? Aw, that's right. Amazon web services, Google Cloud
you know these are all good food, just pick one right and in any great and that's where you kind of start building clearing on, you know, your next most critical piece which is the date a lake, right is in essence, he could be a bit of leg or data warehouse and this is where you put it in like, you know, your customer data, your business data, all the stuff that you maybe have brought from outside third-party vendors, you know, Jason Esquivel, what have you any sort of performance data related to your business? You can all stuff it in there, right? And
that data, will then help you, you know, dude, that model, right? What is machine? Model. AI model. Take from the dealer lake and then you start kind of irritating through those models very, very quickly, right to do what to improve your targeting, your customer, retention to improve your profit and loss, write to contain a risk, you know, there's a lot of stuff that you can do. That model would help you with this should make. And then on top of that is layered, right? You want to lay on, you know, API endpoints with the API
endpoints, you can then share that data, right? With your partner with your customer, you know, with maybe you could potentially in a client's right to a certain insights, that would be helpful and they will be able to unable to businesses. So that's you know, at the end of top the pyramid, you will basically be able to come up with insights, you know, to your business, you know what, no matter what you do, right. And, and that inside is going to be Very defensible. Write you, you can call your company to the next level easily. And so that
I deal with these pyramids structures know, there are layers in there and their modular so they are stuff that you feel is no longer applicable. You can obviously switch them out, right? The modules are interchangeable. The goal is, you want to be use the best-in-class check infrastructure but she kind of gets what's the top you know? That stuff gets very proprietary. You got to have really good data. Scientist data Engineers. Do they handle this right to put that stuff together so that it will continue to derive insights to help you navigate. No difficult times ahead.
Well, that's sort of the meat of the presentation in summary. I would like to know, share quote with everybody here. So this quote came from w, f word. That means he's a famous statistician, a mess magician and professor in a 1919 s. He said In God, We Trust all others bring data. So what does that mean? In Context of 2021? What I think we should continue to stay very optimistic right about Ai and ML and how that would benefit us, right? Sure, there are things. That would be Out of our control and we have to live with that but I think
data should and still is the best lens through which we can understand as a human species how we can navigate and survive and thrive Under The New Normal. That's the end of my presentation. Thank you so much for listening. I'm happy to answer questions if there's anybody who have questions in the steel round. Can you check the current day? Even though we do have one question there, Resay, what you got? A question here, why? When why you mentioned that Landing size bed? You mean, auto loan and housing? Yes. So the Lenny sides
being very, very rough because you know, we we have seen a obviously of rice in default rate and delinquency rate from the boards, right? And so that that's been happening, you know, across different, many categories, doesn't matter with auto, a mortgage or business Landing. I mean in one way or another, you know, businesses or individuals, are they just haven't stopped. I'm paying back because they lost their job or, you know, the circumstances have changed in the life,
right? So what you have seen is in Oprah folios that just kind of going down the drain, you know, polio is a big thing in in the landing space if your portfolio is rubbish, You don't have a business, you know, any types of kind of goes away in landing. And that's why I, you know, I mentioned that and lending has been very challenging year, but I want to say that happy to report in Okemos Financial is going and coming back. You know, we're definitely in the recovery phase. We are not fully out of the woods yet, but we're doing a good job of sort of a rescheduling. Some of these are
longtime payments so that people get some breathing room and make him pay some back as they are recovering their business as well. Cool. I got another question. The Lending Club has Heidi for trait is it higher than Traditional Bank? Well, I don't have that number in front of me but what I can share with you is Lenny Clarke and Traditional Bank. They operate very very differently and my guess is that, you know, probably lendingclub, you know, it's not doing so well because they are the
one that's been very, very adventurous, right? In the way, it is with the lending program. On the other hand, they're so conservative, right? They even have trouble in a hiring the right people that can help them rebuild their attack. So that they can capture the value. You know, of a few people. There are more open to Bank online or bore online. So, And also let me Club. I know, you know, they recently actually require thank, you know, so they are coming in with this challenging Challenger, Bank model and they're so to stop their traditional lending,
because they, they used to do a lot of pure lining. So that kind of I think in in the face of moving data moving away from that. So it's really interesting. You can see that these two are serve colliding. You know, you have bangs that one will rethink text in the European, textile buying a bank. So that, you know, they can operate in a, with a single to of capital to help them move faster. And also, to mitigate risks. Hopefully, that answers the question question. Why do you think Machinery model better than traditional model in lending industry.
Do you see any proof? Well. I want to be completely transparent. There is being there has been Companies that are on different camps. You know, their companies there are totally hundred percent on a i and machine learning when it comes to landing and their companies. On the other side, they're doing more of a roux based approach, right? We are coming up Financial In the middle, but more leaning to the real base. That's kind of what we started with. But now we're leveraging machine anymore. Because what we found that after four or
five years doing Landing, there's only so much that you can do with root base, you just, they're just things that, you know, you're not going to be able to catch with rules, right? But with machine learning, you could do much better job. Of course, you got to go through the exercise of kicking out the right features, right? So that, you know, they're relevant in your model. I saw at the end, we're finding that, you know, we actually have a modeling place and we're sort of right now doing this thing. We are running these two models imperil, right? So that we can see, and compare
which one is doing better, you know, we needed to be, you know, in some ways responsible right. With this way of rolling out these models, so that it doesn't endanger to Cavalia. So we're very excited, we know based on statistics and and some data analysis that was done earlier. If we do the machine learning model, we can improve a lot of things conversion rate before and whatnot. So I really can't, you know, wait for that to happen. When we switch over completely and I think a lot of companies any companies out there where this is
well, Already. So the next question is is Sophia ready to go IPO? You know what, I wish I knew the answer to that question. Definitely put my money on that. You know what's interesting is does being a slew of, you know, fintech companies that have gotten, you know, either Capital through feces or they've gotten Capital by thinking of doing the IPO route obviously. So you know they have been really good in building you know that this community around a product right and they they've done a good job but obviously not
with that. You know there was bumps in a row and I'm sure through covet you know Dave they have there in the Bronx as well. Are they ready to go IPO? I think the landing I feel space is it is still very, very difficult. I think a lot of, you know, investors out there looking out for return on investment, not so much potential growth. You know, I think lending itself is no longer exciting as he used to be maybe three, four, five years ago, maybe. So if I could go in a maybe at some point, so if I was on IPO track but they've kind of
deviated from it, but unique is right now, I think, you know, unless you are like very, very Innovative, a note in your approach and you're branding, and you're probably offering it's hard for you to go IPO else in the US but look at Asia, right? You got in financial that's going IPL fairly soon. Probably you know weeks or months away. You know, and financial is huge and they seen evaluations like 200 billion or something like that and yeah, but you know what the The capture of big big percentage of the market, you know, in China and they have not just
one product, they have like 5 or 16 Tech products or you kind of went place. So yeah, I think it in at the end of day, it really depends on how you position your company, you know, and how you present a strategy advantage to your investors before they can go out with you. Okay, so this is a Jason. First of all, thank you to be our speaker. I'm glad you mentioned the feature. The features I believe Selection is more important than, you know, the motors lektion. That's the key. Even though we know some features,
the indicators for the weekend, not to use them that you know, a lot of things. So you guys trained to somehow, you went to new features or crack multi. How you can handle this challenge. So you know with the You know, that's so let's talk about regulation a little bit. I think Finance space has been one of those, you know, spaces. That's being very tightly monitored, and controlled by the government because of the financial crisis that we had a few years earlier, right? So, you know, since you can't really, you know, use some of those for him
prohibitive features, you just kind of have to stay awake, but that's not to say, you know, you're limited by it. Now. There are, you know, in our analysis and we found features They're good predictors of somebody's willing to pay down the road. Right? And of course I can't get too much into it because he then becomes, you know, proprietary that's the top of a pyramid inside, but what I'm saying is, you know, they are these proximation that you could, you know, you could, you could do
right to figure out if this person is going to be credit with ease down the road. And so that means collecting in a lot of different day. That's right there with the ones, they're not sensitive, right? The one that will not help you discriminate but you're right. But the ones there are no cool by The Regulators so that means you got to have to either right through. You know you can you're going to have to see okay what makes a lot of sense and and I want to say like you know a lot of people in other countries where the financial regulations are not as strict, right? They have obviously
use Some of those more in a sensitive creatures, like gender, are you in line? You can't use gender right? But in other countries, you know, it's whatever. Like so obvious and you can use. The other thing is what's missing in some of those developing countries is that they don't necessarily have a credit bureau, right? So they don't have credit information about this person so they can't even use it for. What they've done is in some of these third-party, developing countries there looking at, you know, telephone bills, right directions to be bills. You
know, what is right? Or even the friends they have on the phone, right? Just to get a sense. Okay, is this guy legit, right? You know, and even things I've heard even things like how often do you call your parents to see if this person is trustworthy? And so there are a lot of ways, you know, to go about this thing and and and I think in order to do it properly in the u.s. in the first of all, you know is I think that company you got to watch the regulation. Do you have to know what you're allowed to do and what you're not allowed? And then start experimenting right with the data. Because like
I said, you know, there's a surgeon data volume, all kinds of data coming in, right? You got to put your food in the house, is picked the right one. And then bill tomorrow. Thank you, thank you for the comments. I would like to go back to previous question from the 10 B about the traditional model. I sign all five years ago Trading Company they were not They were not trust that they didn't trust you no more and I need at least until things change. As I know some of the company Capital One Wells, Fargo Bank of America. Can
you not imagine his surprise discover University catching one piece cover? What do you think make this happened? That's a really interesting question, I think for 5 years ago, you know, there was a lot of hype, you know, there was a lot of hype around by a machine learning. I I remember for five years ago like they were these Robo advisory fintech companies, you know, Tech startups that advice of people, hunting vest, huge, but nobody knew like exactly how they were performed folio eyes. Now, the thing we think tech companies, you know, whatever, whatever
you're doing, abs Financial product, you need to give it time, right? So that you can observe in a performance and see how things do Tramp right at the end of day. And so I think what some of these big Banks or Wall Street firms. I seen that Ashley, they do work, you know, at the end of day they do work, right? So that the thing is, you have to be Comfortable with that. You need to get in with the people. That knows this things right in and out and so that, you know, you can get your fellow hanging
fruit thing that I've seen is also that, you know, some of these tools that are used for building a machine learning models or models, probably just more companies doing it, right? And and so even if you are a vestment formal thing, that didn't really have, you know, the the right profession elizabeta Professionals in house. No, it wouldn't be hard for you to hire a company that could access these tools. Right. That, you know that help you write start these models up and evaluate against existing models. And, you know, that the
other day, when people observe a k, there's a difference, they're going to start, you know, looking into why and how are not. As I mentioned in one of the Spy. Earlier in the day I explain explain an activity is the truck isn't one of the train now, but you can't just build an AI model and not explain what's going on behind the scene. First of all, it doesn't comply with the law. And second of all, you know, people have questions, right? That is actually more risky mauffray. So I think there is going to be more effort in building in AI solutions that would be able
to explain, right? So then people are more sexy in that sense. Thank you. Thank you for you. Call me. I think we, we're so successful for the whole conference. How many speakers are you on for four whole days? Friday, Saturday, Sunday, and Monday? And the last time we talked about in time, thank you. Thank you. Doesn't matter as online or in person, I would love to come every year to go to Los Angeles, Convention Center. Again, Meet the Press impress, a young, professional
entrapreneur, you know, him personally, I really would like to talk with people. Thank you all. Thank you for all the attendees. It won't be easy to work on. Friday Monday and also the weekend and also a special thank you to see Harvey. Thank you to work with me Friday and Monday. It's like cool.
Buy this talk
Buy this video
With ConferenceCast.tv, you get access to our library of the world's best conference talks.