My Career in Data Season 4 Episode 09: Joe Devon, Chair of the GAAD Foundation and Co-founder of Global Accessibility Awareness Day

Hello and welcome. My name is Shannon Kemp and I'm the Chief Digital Officer at Dataversity, and this is My Career in Data, a Dataversity Talks podcast dedicated to learning from those who have careers in data management to understand how they got there and to talk with people who will help make those careers a little bit easier. To keep up to date in the latest in data management education, go to dataversity.net forward slash subscribe.

Today we are joined by Joe Devin, co-founder of Global Accessibility Awareness Day and the chair of the GAD Foundation. And normally this is where a podcast host would read a short bio of the guest, but in this podcast, your bio is what we're here to talk about. Joe, hello and welcome. Thank you, Shannon.

Pleasure to be here. Oh, I'm so grateful that you could be here. Um, you know, we've we were talking briefly uh before the we started recording about how you're kind of data adjacent. You've always been data adjacent, but you're working so hard in data and you do so many cool things.

You know, you've been a keynote speaker at our conferences. Uh what was it? EDW, I think it was uh a couple of years ago. Yeah, yeah, 2024.

Such a great talk. Um, really enjoyed it. And I'm so excited that you could be here and joining us today. Well, you know, I've been, I've been uh, I feel like I've been family with uh Tony and all the conferences for so many years now.

Uh I I love doing anything uh with all of you. Oh, we so appreciate it, Joe. I'm so excited to hear more too about your career and dive in a little bit. So so let's go, let's get into it.

So you are co-founder of Global Accessibility Awareness Day and chair of the foundation that came out of it. So tell me what for those who may not know, what is Global U Accessibility Awareness Day and what is the GAD Foundation supporting it? So Global Accessibility Awareness Day is a day to get uh designers, developers, anybody that builds digital products to become aware that they build products in such a way that it works for people with disabilities. And the GAD Foundation is something that we started after about 10 years of success with the with the day, which was very grassroots, um, and reached we we stopped counting when it reached 220 million people on uh on the hashtag.

Uh so it's a pretty big event. The big tech companies changed the homepage for the day and uh and uh typically and talk about it. And so people started to say we should do this more than one day. We should yeah, we should care about accessibility every day.

And finally we thought, all right, let's do a foundation in order to to help grow this and make it more important all year round. Um and and our mission is about the values, um, sorry, a culture to make sure that accessibility is one of the values that companies care about and to change the culture of digital product development to include accessibility. So when you talk about accessibility, uh I you know it's can you define that a bit more? Go into like some examples uh of what of how you help tech companies make their products more accessible.

Absolutely. There's lots of different kinds of disabilities. The most common one that people talk about uh is if you're blind, uh you need a screen reader because you can't use the mouse. You can use the keyboard, but you can't see where the mouse pointer is, so you can't really use a mouse.

And um, so you have a device called the screen reader that reads the page out to you. And it's really important for you to be able to do keyboard navigation. Like every single uh uh activity that you could do on a on a page or on an app has to be keyboard navigable, right? Which is also great for power users.

Um, but that is an accessibility. Uh I don't even know if feature is the right word. A lot of people get upset when you use the word feature because it's a core functionality. Um, and then you've got audio description.

If you're watching a video and there is sound, that's fine. But you might need to know if you're blind what's going on on the screen. So there is an audio description track uh for broadcast, and it is available if you if the creators provide the the description um on streaming. And then you've got, for example, if you're deaf, you want to have closed captions, or ASL might be your your sign language, might be your your native tongue.

So then you need a signer. Um all of these are accessibility, uh accessibility, just period, without the word feature. I love that you brought into it too the um uh uh you know, if you to using the keyboard, you know, even if you just want to be a power user. I I actually got challenged by uh a mentor of mine um to use Excel for a minimum of two weeks without the mouse.

Well, yeah. It helped me to speed up what I do and how I do it. Uh it was a tremendous experience. Yeah, it it's it's something that a lot of uh hardcore developers get into the habit.

The power users, it's much faster if you use only the keyboard. And um, there are some really cool keyboards out there that unintuitively they reduce the number of uh keys that you press because they have these modifier keys. And what you realize is the less distance your fingers have to run in order to type, the faster you are. So they have modifier keys when you need some of the more, you know, some of the special characters.

Uh so there's a whole uh world of improvement uh along these lines. And when you're making things more accessible to screen reader users, you're also speaking to developers and uh and the kind of people that are a little bit go too far uh in terms of maximizing what they do. Uh and and I'm trying to get there too. So uh I'm one of those idiots uh that that do this.

I'm right there with you. I love it. So so Joe, as the chair and the co-founder, tell me a little bit about your day-to-day job functions. Well, uh, my problem is I come up with too many ideas and then I wind up like doing a ton of projects.

And now with AI, um, I saw a great quote that said that uh I used to have two to three side projects that never got finished. And now with AI, I'm able to do 10 to 15 side projects that never get finished. Um I love that. But I didn't really answer your question.

So uh day-to-day. Uh one of the big things that I'm working on is a benchmark for accessibility. So AI models, they the coding-assisted AI models, they're writing the code today um for you. It's gonna, it's in the middle of taking over development everywhere, uh, but how accessible they generate their code is very important.

And so I created a benchmark under the auspices of the foundation in partnership with ServiceNow. And what we do is we test the HTML that AI models come out with, and we put it into a leaderboard. So it's a benchmark that says how well these models write code. But there's a lot of other benchmarks that we're going to come out with uh along those lines.

Uh and it's called AMAC, and it's at aimac.ai. Oh, very cool. Uh, I love that a lot.

So um, so how are you working with data in your job? Well, we generate every time we run one of these benchmarks, that by itself generates a thousand different web pages that also generates uh a lot of um uh uh JSON reports that explain or that basically score um all of the web pages that are generated. Uh there's the and then there's the HTML uh directly, there is the the prompts that go with it. There this by itself is just a good a good chunk of data.

Uh, but more importantly, where uh we're talking to different organizations, the data that trains the models is really important because uh let me give you one example. If you are trying to do sign language, there is a fair bit of of data on American Sign Language, but there are as many sign languages as there are languages. So you have Japanese Sign Language, British Sign Language. Um at the Super Bowl, the some folks um some signers got upset because the uh it wasn't ASL, it was Puerto Rican Sign Language.

Um and so some people got upset that it wasn't in English because in even in Puerto Rico, most people uh use ASL. Um so just an example of some of the the details in the accessibility community. But if you want the AI models to understand it, you have to have all of this data from people with disabilities and how they communicate in order that the models will be able to handle it. That's fascinating.

Uh and things that you you wouldn't think about um on a day-to-day basis. Uh it's a lot of data. All right. Well, Joe, let's talk about how you got to be into this role and founders of uh the founder of you know such important um topic.

So tell me, is this what you wanted to be when you grew up? When you were saying six years old, did you say I'm gonna grow up and be a co-founder of uh Global Accessibility Awareness Day and and a founder of the foundation? No, people would tell me that um I should become a lawyer uh because I would advocate a lot for my brother, for example. Um, but uh no, I wanted to be a hockey player, and I'm still upset that I never had the talent to be a hockey player, but that would have been nice.

I love that. Uh you're the first uh hockey player dreamer that that I've had on the show. I love that. Thank you.

Um so but then as you were um, and I love the reference to the lawyer. Um so as you're growing up, then as you're you know, starting to select your classes in school, and and what are you leaning towards and where do your passions start to grow? Well, by the time I was 13, I got an Apple IIe at home, and um I had no access to books. There was no internet back then.

All I saw was this black screen, and I didn't know what to do with it, but at school we had a bunch of people that had a whole bunch of discs with uh floppy disks, the five and a five, what is it, five and a quarter inch um floppy disks. And we might have been doing a little bit of piracy and and passing along uh some programs. So that that was kind of interesting to learn the different things going on. And um then within about a month, I wrote sort of a crude Hebrew translation uh application um and just kind of figured it out.

But it was it was a lot harder. Now you can actually read instructions, manuals, and things like that. Um, so I took to it at a very young age, but um then I don't know, in school, I I I had um I skipped a grade and then I went early admissions to school. So I started at 16, which was too young.

Um, I just wanted to have fun and I didn't really pay enough attention in school. Uh I did take some computer classes and uh and was bored from them, so I didn't pay that much attention. Um and then wound up doing uh trading, international trading, which is basically import-export, uh, in the 90s, just as Russia was opening up. So I really had like that business mind and uh and had some really crazy adventures all over Eastern Europe.

Um, I could tell you more, but you know, um you might want to comment so far so that I don't uh just ramble on too much. No, it's it's fascinating. So uh, you know, as you're you're traveling, you know, what are you doing for for work and and how are you you know managing that? Yeah, so what was going on there was just um Eastern Europe, you know, it was Russia, the Cold War was ending, and they had massive need for for food.

Um, and that was mostly food products. So I sourced out the food from factories. I was living in Belgium, sourced out the food from factories, and then uh shipped it into Russia, sold, sold by the container load. Um, but it was the reason I say it was a wild is because we had um trucks that were getting shot up on the way over there.

And it was very difficult to get to to get it there securely. Insurance was really tricky. Um and the I I became friends with some freight forwarders that said that their drivers would come back every two weeks, um, you know, uh basically shot and and killed. So it was kind of the Wild West.

And uh as much fun as I had, I had to kind of give that up after a while because it was just clearly too dangerous. I I had friends who uh had one friend who was in a plane crash um to Romania, just some some rate, you know, wild events and you get threatened uh if your goods don't arrive, that kind of thing. So that was kind of a total detour. It wasn't related to data at all.

I'm sorry. Okay, no, that's that's what it's all about to hear the path, right? Yeah, to where you got. Um that's fascinating.

Uh wow. Um okay, so you you really are just like an entrepreneur and a go-getter from the beginning. Yeah, yeah, uh, very much so. And then uh came back to America.

My cousin was working for a search engine company that was around before Alta Vista, let alone uh Google, uh called CompassWare. And I had no idea what a search engine was, nor did most of the world, but he said this, you know, I was using CompuServe a lot since since I was 13. Uh, and he said, if you think CompuServe is good, the internet and the world wide web is insane. And uh and so he brought me along there.

Unfortunately, the search engine company didn't last that long. Um, but then I joined a consulting company uh called Predictive Systems, and uh we took it, it was an early employee and we took it public uh four years later. So that was uh that was good. It was one of the first few uh agencies.

And then from there, um I took a little bit of time off, but then uh um American Idol uh called. I got a phone call from a recruiter uh asking me what kind of work I'd like to do. And I was like, uh, I think performance is uh is gonna be, it's just interesting to me. Like, can you build a system that's really performant?

Uh and I said uh American Idol would be one of those examples where there's a ton of traffic. You know, it was the number one show on TV at the time. More people voted for it than for the president. Um, so I said uh so I said that, and the recruiter said, Oh, well, I'm calling you for American Idol.

Uh and then uh that was a Friday. I packed all my stuff up into storage. Uh and by Monday, I was in Los Angeles and working uh on American Idol. So that was a pretty crazy path.

Yeah. Um so what we so what did you do for them? I was a back-end programmer. So I helped them um with their databases, with their their data, um make sure that everything stays up with the performance.

And then uh we had a crack team. Uh after a couple of years, they were very happy with what we did and they said, all right, we'll trust you to build the voting app. So then we built the the first like real major voting app for a reality show, and it went great. It went off without a hitch.

The one day behind the scenes, uh all of the so they pulled more and more people off of it because they they had all this trust, and in the end, there were no back-end developers. So two weeks before we go live with the voting, they call me in and they're like, we need to have a look at this. And so uh we were about to go live with the voting. We were three minutes to go, and all of the servers were completely maxed out.

And I had to go in there and uh and look at the code. I found it, the problem, and then yelled, bounce all the servers. I yelled around the corner, bounce all the servers. So with three minutes to go, we were maxed out, two minutes to go, all the servers were at zero.

And that's when the gray hair kind of started because I knew we'd be on CNN and and all those shows. Uh so yeah, just a bit of a crazy, crazy story. Very, very cool though. Um, I mean, because it it was one of the first shows.

I mean, it was the first show with you know the voting. And so, you know, it to be on the front lines of that, that's uh uh very cool. So, where do you go from there? So then from there, uh because we had all of these challenging uh situations with uh scale, uh, and I had in New York, I was a part of a lot of tech meetups, and there was nothing in LA.

I started seven different meetups um around performance and a variety of topics. Any any technology I was touching, I started a meetup and uh kind of got well known in the LA ecosystem. And that's where I met uh Tony uh Shaw and uh Semantic Web. What do we we called it something like that, the some the semantic meetup or something like that?

Um and uh then from there I you know I got to know people and then finally started my own agency uh and got a lot of business, including from you know American Idol and and uh lots of other reality shows, and uh grew the company to a hundred developers. Um and we're doing great. We were gonna double in size in 2019, and then COVID hit, and uh what do you what can I say? From there, the the company went down and uh and I had to reinvent myself again.

And uh and that's where that's kind of what got us uh here. Just for reference, in case any of our listeners don't know, Tony Shaw is the founder of Dataversity, um CEO and founder. So he's he's great. And I met up with Tony just as he was uh selling Symantic Web.

So it's been it's been a while. Yeah. This is, I mean, it's so fascinating, Joe. I mean, you led a very colorful uh and and very fun uh career here.

We get this question a lot in our webinars. How do I prove my data governance skills? Dataversity's applied data governance certification. It turns what you're learning in our free sessions into a credential that actually signals expertise.

If you're leading or launching data governance, search applied data governance certification from Dataversity and see what's inside. So our 2020 happens, your current uh company is not thriving. So, but what made you land then on global accessibility awareness day? Why, why was that the next step?

Well, so so I kind of started I started my uh my agency slash you know dev shop around the same time as I wrote a blog post proposing global accessibility awareness day. My dad had been um trying to bank, he was 87 uh and spoke 11 languages, uh survived the concentration camps and was just a man that uh when he walked in the room, he commanded respect. You could just there was just something about my dad that was special, and you could everybody saw it instantly. Um and so watching my dad be unable to bank because the bank's website was inaccessible was really painful.

Uh so I I had a blog called MySQLtalk.com, which had maybe 10, you know, 10 devout subscribers, that's about it, and wrote a blog post on there proposing this Global Accessibility Awareness Day. And to my utter shock, the idea went viral right from the beginning. We were in 16 cities in the very first year, and now uh we've run publicly over 2,000 events.

When I say we, I mean the community. They you know, just the community picked this up. Um, so this happens globally. uh every year and there's at least as many private events by corporations that want to improve their accessibility internally.

Wow. That's quite the story and quite the uh motivation. Yeah. And then and then this thing is just I mean so many, so many events uh happened around it that um you know we we had um uh Stevie Wonder did a concert for GAD uh uh at the Apple campus um we've had Sir Tim Berners Lee co-keynote with my co-founder uh and then in 2024 I get a call you know usually I do about 10 keynotes on that day in 2024 two weeks uh before the day I get a call from the White House that they're running a GAD event and can I come over and speak?

So this thing went uh just in 13 years went went to the White House. It was pretty pretty uh I don't know I don't even have the words for it. It shows you you found uh you know a very big need. That's for sure.

I mean people there's it people always struggle to get a budget uh for accessibility but if you're wondering how important it is and how many people it impacts you can just look to a blog post that reached at first 10 people and then now you know reaches a country worth of people every year. It's it's I mean the the numbers speak for themselves right absolutely tell me you know what in in this very impressive career what's been your biggest lesson so far? Oh um that's a good question uh there are so many lessons uh um you know as an entrepreneur you have to take risks but uh people don't always understand that that risk does not mean it's real um and if you're an if you're a true entrepreneur you cannot see that a risk means that you might fail you you just can't imagine that you're gonna fail you for sure you go through the tough times and you know you will um but you just you figure if you don't give up then you won't fail and I still think of it that way because you know thank God I'm doing lots of interesting things and uh and and uh covering what I need to um but it's it's really real like you there's no guarantee if you're an entrepreneur that you're gonna you know succeed on the other end and even though I was right there my retirement was right there um you get COVID out of nowhere and it changes the entire market underneath you um so yeah so I'd say uh really think about what you do and try and get yourself that like I could have exited many times over the years I should have taken that first exit and then had that security and then you know do a new one.

So I'd say that's the biggest thing that I've learned. And now with AI uh changing the game completely I think that people absolutely need to think about uh what can they build that's going to uh be sustainable. And and I think that it's a data story. Uh it it really is a data story because software costs nothing to create right now.

You can make five copies of it you can copy anything you might have a piece of software that you've um you think you have IP you can keep it closed source but as long as it's available to anybody with uh AI coding assistant they just point at it tell you all the features and it will recreate it uh very quickly so where do you have the ability to have a business that will sustain itself is if you have data expertise in a particular domain and data around that domain. And if you don't then your margins you might still be able to stay in business but your margins are going to be compressed because the barrier to entry goes down and your data is your barrier to entry right so I would say uh this is this is something for leaders in this space to really think about and you know the the audience your audience here is in a special place because if you are a data expert you can understand how to implement AI in a way that you have some kind of mode very very true.

Yeah it's it's been interesting to watch companies uh working to implement AI without uh any kind of data management and then coming going oops wait yeah yeah for sure and and it and is people are in a really bad position because if you don't implement AI then you're gonna fall behind. You absolutely have to get your people up to date and and anybody that's programming and not using AI uh like they're they're in for a world of hurt they you know and and if you're not allowing them in your organization to play with it then you know the best people are going to leave.

So you have to implement it but it's not really secure yet. So uh it's it's sort of a catch 22. Uh it's fascinating times but it's it's tough times too. It is indeed um time of opportunity for sure so uh so don't you know I mean as diverse as your uh businesses have been what is uh you've been in data like you say for a long time working with data um throughout most of your career what is your definition of data my definition of data um that's that's a good question um I know it when I see it you know um data is just a symbolic form of symbolic representation of anything really um in an organization it's a symbolic representation of the domain specific um information that that you have in your company it can be as simple as um as the laws that your organization is specialized in helping other companies deal with so for example in HR if you've got a a people company that that helps handle people you've got um you've got all this information about the particular employees so how do you handle um the the permissions in order to access it how do you follow the laws all of that is codified um with symbols uh uh of your information and that is data yeah does that work yeah absolutely it's very nice so so do you see the importance of data management and the number of jobs working with data increasing or decreasing over the next 10 years and why 10 years I mean what's gonna what's it gonna look like in one year is the question but I think if anybody has um has some stability in terms of their job it would be data because there's gonna be so much more data first of all the data is used to train the models it's so so important um uh as the foundation of these large language models then you also have structure data right now the large language models are very non-deterministic it's all probabilistic and you need to take this probabilistic information and turn it into some kind of structured data and even though the LLMs themselves can help you with doing that you still have to put certain guardrails around it and it hallucinates a lot.

And so if you have uh the know-how on the data side to direct it to set up your systems I think that that those jobs are about as secure as any that are out there. But again like well what's going on these days with these agents and these swarms of agents uh I I mean I don't know who's who's secure anymore I really don't know right it's yeah it's it's changing so rapidly um I agree uh what advice then would you give to people looking to get into a career in data management to get into this space uh really reading read a lot of the papers the the most interesting papers that are coming out and make sure that you understand how LLMs work it's it's sort of a miracle that LLMs work at all but I'd say that 99.

9% of the people that use them have no idea how they're trained and it's a fascinating story it's a data story it's like really hardcore a data story and if you understand it you're gonna be very well placed to use it and to control uh AI. Yeah very much so oh well Joe um it has been such a pleasure chatting with you today and hearing more of your amazing story and the the places that you've been and the things that you've done. I'd be remiss if I didn't uh ask, you know, when is Global Accessibility Awareness Day and how would people find the um GAD Foundation?

Global Accessibility Awareness Day it's easy to figure out the date every year because it is the third Thursday every May. It's the 21st this year um and the way that you find uh the the site itself is accessibility.day so that's a c e s i b-i-l-i-t-y dot day uh and then the GAD foundation is g-a-a-d dot foundation um that that's how you reach it I love those URLs very very nice yeah oh well Joe it has been such a pleasure chatting with you likewise I'm glad we finally got out got uh to do this show uh yeah uh me too and uh so impressed uh I could probably talk to you for hours but yeah so but to all of our listeners out there if you'd like to keep up to date in the latest podcast and the latest in data management education you may go to dataversity.