Rehgan Avon - Applying Artificial Intelligence to Strategic Initiatives

The conversation this week is with Rehgan Avon. Rehgan is co-founder and CEO at Align AI. Align AI helps AI experts scale their knowledge faster and more effectively within their organization. She was the founder and is now a board member of Women in Analytics, whose mission is to increase the visibility of women making an impact in the analytics space. And provide a platform for women to lead conversations for the advancement of analytical research development and applications. And if that wasn't enough to keep her busy, she is also on the advisory board of OhioX.

If you are interested in learning about how AI is being applied across multiple industries, be sure to join us at a future AppliedAI Monthly meetup and help support us so we can make future Emerging Technologies North non-profit events!

• Emerging Technologies North
• AppliedAI Meetup

Resources and Topics Mentioned in this Episode

• Align AI
• Women in Analytics
• OhioX
• Natural language processing
• Amazon QuickSight
• DataConnect Conference
• Renée Cummings
• Chip Huyen
• Cassie Kozyrkov
• Noelle Silver

Enjoy!

Your host,
Justin Grammens

Transcript

[00:00:00] Justin Grammens: Greetings, Applied AI Podcast listeners. This is Justin Grammens, your host of the Conversations on Applied AI Podcast. Just dropping in to let you know about a very special event we have coming up on Friday, November 10th. It's the Fall 2023 Applied AI Conference. You can learn more by going to AppliedAIConf.com. This full day in person conference is the only and largest artificial intelligence conference held in the upper Midwest. It will be in Minneapolis, Minnesota on November 10th. We will have more than 20 speakers with two tracks covering everything from AI, business applications, chat GPT, computer vision, machine learning, and so much more.

And for being a listener to this podcast, use the promo code of podcast when purchasing your ticket for a 20 percent discount. So here are the details. Go to AppliedAIConf. com. That's AppliedAIConf. com to see the full schedule and register for the only and largest artificial intelligence conference in the upper Midwest on November 10th.

And don't forget to use the promo code of PODCAST when checking out to receive your discount. We look forward to seeing you there, and thank you so much for listening. And now, on with this episode.

[00:01:02] Rehgan Avon: I think for a long time it was a bottom up initiative, right? We had people who were trying to find feasible use cases with the data that they had.

And so maybe those weren't strategically interesting to a lot of organizations. Um, but they were doable. I think what has changed is this perception of what AI can do. I mean, if you think about it, Chad, GBT gave people the ability to interface with an AI system in a deeper level than people are used to interfacing with.

[00:01:36] AI Voice: Welcome to the Conversations on Applied AI podcast where Justin Grammens and the team at Emerging Technologies North talk with experts in the fields of artificial intelligence and deep learning. In each episode, we cut through the hype and dive into how these technologies are being applied to real-world problems today.

We hope that you find this episode educational and applicable to your industry and connect with us to learn more about our organization at AppliedAI. mn. Enjoy!

[00:02:06] Justin Grammens: Welcome, everyone, to the Conversations on Applied AI podcast. Today, we're talking with Regan Avon. Regan is co-founder and CEO at Align AI. Align AI helps AI experts scale their knowledge faster and more effectively within their organization.

She was the founder and now a board member of Women in Analytics, where their mission is to increase visibility to the women making an impact in the analytics space. And provide a platform for women to lead conversations for the advancement of analytical research development and applications. And if that wasn't enough to keep her busy, she is also on the advisory board of OhioX.

So thank you, Regan, for being on the program today. Thank you. I'm so

[00:02:42] Rehgan Avon: excited to be on.

[00:02:43] Justin Grammens: Awesome. Well, I gave a quick synopsis of where you are today. Maybe you could give us a brief background in regards to how you got to where you are today throughout your career. Yeah, would

[00:02:51] Rehgan Avon: love to. So traditionally, I have an engineering degree from the Ohio State University, and I have to say it that way, or else I'll get booed.

So I've been kind of in the data science space for about 10 years and specifically enterprise AI for the last eight. So helping organizations design, build, deploy, manage, monitor machine learning systems for a number of years. And I've really, really enjoyed that. That brought me to where I am today, which is Align AI.

My co founder, Brendan, and I started about three years ago. Basically, like you said, to help companies kind of build, adopt, and govern their policies and procedures around data and AI, which, you know, we're big process nerds. Not everyone is, but we love it. We became pretty obsessed with this idea of this kind of like AI adoption problem, getting people to.

Actually adopt solutions and see benefit out of it because for a long time, we saw a lot of research, unfortunately, go nowhere and people didn't actually realize the value of what they were trying to do. So, you know, we started the company to address that and then along the way, built Women in Analytics to help get more diversity into the space, host a conference every year.

So, yeah, it keeps him busy

[00:03:59] Justin Grammens: for sure, for sure. Well, you've been in this enterprise AI space for a number of years. How have you seen it change? Right? I can imagine. I mean, AI really wasn't the big buzzword. I feel like, you know, five or six years ago than it is today.

[00:04:12] Rehgan Avon: Yeah, it's interesting. I think you get a little bit of bias being in industry for a decent amount of time where you're like, Oh, yeah, yeah, you know, everybody's been going after this kind of like AI enterprise AI space for a while now.

But the reality is that we're at a whole different level. We've never had more. C level individuals ask questions about how they could leverage AI strategically at their company. We talk to a lot of legacy enterprises or traditional enterprises. So I'm just speaking in context to those types of organizations.

But, you know, we've got. board initiatives on what's our strategy with AI? How are we implementing it? How are we leveraging it to be competitive? And we just didn't have those types of conversations before. It was a couple of folks in the technical teams who were coming up with some use cases and working to kind of operationalize some more simple examples or use cases.

And that was always fun and fascinating and interesting. And we sifted through all the technical nuances to get that done. Um, but now we're talking more strategic. And so I've seen the conversation shift more strategically. I've seen the budgets open up a little bit. I've seen reallocation of research and innovation budgets, uh, point towards these AI pilots and use cases.

And so, yeah, it's been, it's been super fun and a crazy time over the last eight months. Uh,

[00:05:28] Justin Grammens: cool. Yeah. I think I heard some stat, right? It's like 70 percent of AI initiatives never actually see the light of day, right? Is that true? Yeah,

[00:05:36] Rehgan Avon: definitely. It actually might be worse than that. Yeah. Yeah.

[00:05:39] Justin Grammens: And I saw the same thing.

I've been working in the Internet of Things for many, many years. So a lot of companies talking about wanting to add sensors and that's kind of how I got into AI was I started building a lot of connected products with companies. And then once they started getting all this data in real time, they're like, we really need to start putting algorithms around what's going on here.

It's just too much information. But I was doing IOT 10 years ago, you know, now before it was really even called Internet of Things. But I saw the same thing. I saw companies, you know, wanting to do this stuff, but then not really investing a whole lot in it outside of just some simple R& D groups and them saying, well, let's just try a little pilot over here.

But then there was never really some sort of corporate buy in with it. But it feels like AI has sort of, you know, crossed that chasm in some ways. You know, there's probably a number of reasons as to why do you think, I guess, we're sort of seeing all of this interest in it

[00:06:25] Rehgan Avon: right now? Yeah, I think for a long time it was a bottom up initiative, right?

We had people who were trying to find feasible use cases. And so maybe those weren't strategically interesting to a lot of organizations, but they were doable. And so, I think what has changed is this perception of what AI can do. I mean, if you think about it, ChadGBT gave people the ability to interface with an AI system in a deeper level than people are used to interfacing with them.

So, for example, we all interface with You know, technologies like what Netflix has built or social media, like we're interfacing with AI all day long. We just aren't doing it consciously and ChadGBT gave us the ability to kind of play around with it. And so I think it opened up the minds and gave this like paradigm shift to a lot of individuals who have great ideas on use cases, but maybe didn't understand the nuances of AI or the power of AI.

forehand. And so that's when we started getting more interest. People started coming up with more interesting use cases and more impactful use cases. And I think the vision was there for a lot of people who are able to point budgets in the right direction. So that to me and plus all the fun marketing hype that comes with it too.

So there's some FOMO happening, but yeah, it's definitely an interesting time for it.

[00:07:48] Justin Grammens: Yeah. What are some interesting applications, I guess, when you go into some of these businesses to sort of say, you should be thinking about AI in this, in this way, or have you, or have you thought about it in this

[00:07:57] Rehgan Avon: way? So it really depends on the type of company, you know, for companies that have been, building models before they understand, or at least their technical teams understand what AI can do.

And then, you know, for individuals who haven't been a part of the conversation beforehand, we often have to do a little bit of educating around what is AI and what's the difference between You know, the more traditional techniques, umbrella and enter AI, like the traditional machine learning techniques that people have been using for a long time and generative.

What is the difference between those and, you know, generative is a, we're talking about like multimodal models or language models, utilizing text content and things like that. And that is different from what we've traditionally used in the past. For the most part, we've done NLP for a long time. So I think it's just like, number one, it's just orienting people to what is this whole ecosystem of AI, not just the chat GPT stuff, but everything else.

And that usually sparks conversations on what they've already been doing, you know, their data science teams have been working on versus, okay, yeah, we do want to pilot some of these more content heavy use cases with generative, for example.

[00:09:04] Justin Grammens: Yeah. Is it kind of 50 50 where companies have already sort of dipped their toes in the space and others completely newbie to this whole thing, or are you seeing more companies adopt it that have already sort of tried stuff, or I'm just kind of curious what you're seeing with regards to the

[00:09:17] Rehgan Avon: breakdown.

Yeah, so as far as generative goes, I think everybody's kind of dipping their toes in for the first time. And most enterprises are just now getting to the point where they've got maybe a defined pilot that they want to try. Maybe they have identified some data that they want to work with. They're still having tons of conversations with security and governance teams and legal and risk and compliance.

Like, what can we do? What can't we do? What are the guardrails we want to put in place? How can we do a kind of risk averse type pilot just to see what's possible? And so. I've seen companies in iteration mode of that. So they're having strategic conversations around that, typically with their innovation teams, plus maybe a couple of stakeholders.

I've not seen many companies operationalize these things, put them into production and, you know, they've found something super great and then they're moving forward with it. Not yet, at least. And then on the more of the traditional techniques, for the organizations who haven't been doing this, they have a big journey ahead of them.

And I don't think they understood that beforehand, right? Data quality issues, data accessibility issues, you know, if they haven't already started laying a foundation for that, they're in for a lot of work. And for the companies that have been, they get, they get that. And so, they maybe have better curated data sets, better quality data sets, and this idea of piloting something in generative AI is a little more obtainable and powerful.

[00:10:37] Justin Grammens: Yeah, yeah, yeah, no, I, I mean, I, I kind of do a similar thing because I do a lot of AI consulting as well, and I really try and have companies focus on what's some low hanging fruit, what's a small piece of data or a subset or a very small use case, like, and sort of start there rather than sort of blue sky on like, let's boil the ocean and do everything.

I'm assuming you kind of take that same philosophy as you come in.

[00:10:59] Rehgan Avon: Definitely. Low risk use cases are the best ones to start

[00:11:02] Justin Grammens: with. Yeah, and I guess that kind of goes back to the 70 percent that we're talking about in regards to the failure. You know, there's a lot of reasons as to why they either fail or they don't move forward.

You know, sometimes it's over promise of the technology, you can't actually deliver it. But also, and I think one of the other things is just not clear guidance on it. Like, what is the eventual outcome? Right. And even like, what is the goal line at the end of this? A lot of things, a lot of projects can just sort of churn and churn and churn, and you never end up actually saying.

Hey, have we accomplished anything or not?

[00:11:29] Rehgan Avon: Absolutely. Yeah. The companies that usually start with it as a research project, those aren't the challenging ones because you're almost a hammer looking for a nail, like, Hey, we found something interesting, but we have no idea who's going to own this, who's going to operationalize it, what sorts of processes is going to change, what.

sort of risk tolerance we have with this thing. You know, what sorts of stopgaps we have in place in case this starts not performing well. Like, it's a really scary thing to have a series of dependencies on business operations and then just say, now we're fully dependent on this model and if it goes down, we're in trouble.

Or if it stops working, we have to kind of default back to this other operating model that we were doing before and that's a scary thing for organizations. So, I agree. I think there's a risk tolerance associated with that. But also, you know, for a long time, we've been doing a bottoms up approach of modeling and not a top down approach of modeling and that's created an ownership problems.

[00:12:25] Justin Grammens: For sure, for sure. Um, you mentioned some of the barriers as well. You talk about legal and other aspects around like data. I mean, for certain industries that are very highly regulated, you're seeing some slowdowns or some apprehension with regards to not being able just to dive in and start, start working.

Yeah,

[00:12:41] Rehgan Avon: we work almost exclusively in. You know, highly regulated ecosystems. So we work in health care and banking and insurance and, you know, automotive and manufacturing. And those industries, they do move a little bit slower. They are a little more apprehensive to trying some of these more interesting use cases that might have a sensitivity element around it.

They are a little risk averse in the sense that. You know, they don't want more regulators coming to them and asking them questions. They're already spending a lot of time on data privacy and data governance to address that. And I think in addition to that, they are waiting for somebody else to set the guideline so that they don't have to be the ones to do that.

So that could be government, that could be You know, the regulatory bodies and I think there's still a lot of debate and discussion around what that looks like and a lot of laws getting passed and a lot of state level, country level that is still somewhat unclear and also very complicated to navigate.

So I think there's hesitancy because of that. And it's not that they're not moving forward. They're just not selecting certain use cases that they know will have a little bit more of a overhead on the regulation

[00:13:59] Justin Grammens: side. Gotcha. Sure. And in a lot of cases, I mean, I feel like they have the data because they've been doing traditional data science, you know, they've been doing predictions and predictive models and stuff like that.

And a lot of these situations, is it the fact that now this is going to be maybe loaded into an open AI or my data might be exposed to somebody else? Are those some of the major concerns, I guess? 100%.

[00:14:20] Rehgan Avon: Definitely some of the major concerns. I think, you know, nobody wants to be the first one where their data is leaked or, you know, their data is now incorporated into a model that other people can leverage, you know, especially banks.

They've been burnt by that multiple times by vendors who leverage data. to train their models, and they're not willing to do that again. And so, I think there is a lot of hesitancy. Obviously, like, OpenAI is trying to address that with some of their enterprise features and functionality, but there's still a big, at least from what I've heard, is a big build versus buy.

Like, do we go the open source route? Do we, you know, do we build out kind of our own enclosed architecture? Is it worth doing right now for like one use case or a couple of use cases? I think some of them are leaning on Azure. quite a bit to support them. And so there's still not a lot of direction there on what companies have decided, but it's definitely a topic of conversation.

In fact, we're doing this big workshop in two days with a bunch of CISOs, you know, chief information security officers, just talking about how you set up productive frameworks internally to continue iterating and innovating, but also keep You know, keep data safe and keep the company safe. So it's a dance for sure.

I just think people are still trying to figure out what that is or what that

[00:15:37] Justin Grammens: looks like. Sure. Sure. Yeah. And any, any new technology, I guess, can be a little bit worrisome to businesses in this particular, and I guess to bring it back to some of the IOT stuff that I have worked on, you know, a lot of these companies are already profitable.

So they're kind of like looking at this thing saying, well, this is what we don't need to add this stuff to our existing hardware. This is the way we've always done it. You know, we've, we've always had these machines that just run for 40 years or, you know, uh, or we've always had these systems out in the field and we just drive around and check on them, right?

So a lot of times it's like, there's just, you're kind of fighting this in, this sort of inherent of like, this is the way the business has always been done. So why, why rock the boat?

[00:16:13] Rehgan Avon: Oh, a million percent. I have always said these enterprises like, They've been a bank for 50, 60 years, and they've been fine being a bank, you know, they didn't have to be a technology company.

Now I think people are feeling more pressure that they do. And that's, you know, that's been built up pressure over time. But I think the biggest piece is like that defensibility, like we've been a bank for a long time, we've been in the market for a long time, we have a lot of really good data, you know, and data is really what's valuable here.

And how do we leverage that? And I think they're starting to get more tolerance on. experimenting with some really inefficient areas of the organization, but to fully replace like a core operational element of what they're doing, I think, yeah, it's just really hard. It's hard to justify that big of a

[00:16:59] Justin Grammens: shift.

Yeah, yeah, for sure. Because yeah, it touches all aspects of the business. And that's the beauty of AI really, is that it can be applied in so many different areas. That's kind of what has drawn me to it over the years. It's that it's sort of, people call it a general purpose technology, right? So it can be used across all sorts of other purposes within your organization from finance to HR to product development, you know, to coding, right?

It can start, it can write code now for you, right? So there's just, there's like really no area that AI can't touch. One of the reasons as to why I started this podcast and everything.

[00:17:31] Rehgan Avon: Yeah, it's fascinating. And I think one of the most fascinating things data did for companies, as at least AI is driving this even more, is this cross sectional element of like operations data, plus HR data, plus financial data.

You know, these models, the more context you give them, typically the better they perform. And these organizations are not used to sharing their data that way, joining their data that way. combining their data that way. And sure, we've tried it with dashboards, but that's still been fairly siloed as well.

But AI, you know, it needs that context, so it's been really fascinating watching cultural shifts being triggered by these types of initiatives as well, where you have these people who aren't used to working with each other. But need to for these different initiatives. So it's kind of fun.

[00:18:17] Justin Grammens: Yeah. I think the whole chat side of it, the whole language side of it is just fascinating to be able to, you know, use a tool and just sort of converse with it.

Yes. You could always get a dashboard and maybe drill into some data and write a query or whatever it is, but just. The whole fact of just being able to sort of talk to these things, I think, is sort of, has really blown people's minds.

[00:18:36] Rehgan Avon: I couldn't agree more. I'm so excited to see where, kind of, business intelligence goes.

Like, we're just so in the early phases of this shift. You know, we started seeing, uh, I think it was QuickSights by AWS. They started creating, they incorporated, like, natural language processing into their tool a couple of years ago, actually. And it was, like, pretty good. And so, you know, just to watch where we are going to be able to go, where it's like, I can just have a conversation with data at a company.

I think that's such an interesting vision that a lot of organizations are pursuing.

[00:19:11] Justin Grammens: Yeah. Yeah. Great. Well, we've spent, you know, 20 minutes here or so sort of talking about AI, which is, you're doing a lot of that today at Align AI. But I did, I did want to talk a little bit about the Women in Analytics mission, right?

The work that you're kind of, maybe discuss a little bit around sort of like what made you want to, in some ways you're kind of giving back, right? You've, you've been successful in this space, right? And now you decided to form this organization a number of years ago, actually, to sort of help other women.

Kind of get more and more into this analytics. Maybe you could talk a little bit about the history of that. Yeah.

[00:19:41] Rehgan Avon: So I actually started the organization, gosh, probably like eight years ago. And it was out of necessity of not being able to find other women in the space. You know, in academia, there was actually a lot of women in stats and math.

And I looked into it a ton because at the time I was still going to school. And, and then I started joining a lot of industry groups and organizations, and that's when I noticed the big difference of like, there are not a lot of women showing up to these events or that have presence in conferences, you know, some somewhat research papers, books as authors.

And to me. you know, looking at the data, there are tons of women who have been actuarial scientists and statisticians. And I think when the space, the data science ecosystem got more computationally heavy, more C. S. Harvey, we started seeing kind of that applied elements of machine learning in industry.

You saw less and less women just because in computer science, there's a lot less women. And so for me, I just wanted to be able to provide visibility to those women because I was looking for them and couldn't find them. And so I was like, well, maybe if I create kind of this center of gravity where women want to come and, you know, talk about what they're working on.

You know, maybe we can create our own network and create that network effect. And so that was the goal and the objective. It was to provide opportunities at a large scale for women to share their work. And it could be in academia, it could be an industry. And I just wanted to provide that platform, help.

More women get published, more women write books, more women speak at conferences. And so that's what we did. And it's been great. I've met like some of the most incredible people through the organization and it continues to kind of grow at a global scale, which has been so fun to

[00:21:29] Justin Grammens: watch. That's awesome.

Yeah. So it was started, you know, by you and people in Ohio. Is that

[00:21:35] Rehgan Avon: right? Yeah, kind of as a grassroots effort in Ohio, of all places, right? Oh, there's a

[00:21:42] Justin Grammens: lot of interesting things going

[00:21:43] Rehgan Avon: on in Ohio, but yeah. Oh yeah, yeah, there are, there totally are. And then, you know, quickly kind of scaled beyond that. I'd say like, maybe at this point, 60 to 70 percent of our speakers every year at the Data Connect Conference are from outside of Ohio.

So we bring in kind of experts in different areas from all over the world into Ohio every year and just the amount Like we survey all the speakers afterwards, like, did you make business connections? And did you make meaningful career connections amongst the other speakers? And we get such incredible results back every year, which is so great.

[00:22:16] Justin Grammens: That's awesome. Yeah. How did you handle COVID? Just went online.

[00:22:20] Rehgan Avon: Yeah, we had been running, we had two major conferences before COVID hit. We were on our third. And so we canceled in 2020. We thought about going virtual, but it was two months away. So we were like, I saw the platform is trying to, yes, everyone has virtual events.

And I was like, I'm not going to experiment with that. So we decided, listen, we'll put a pause on this. We launched our membership platform where people could still connect virtually and, you know, get mentorship and things like that throughout the year. And then we slow launched again, did a smaller event in 2021 and then kind of came back from that.

So it was a small build back, but yeah, we, we grew our virtual presence a ton and, and we've been doing hybrid ever since, which I actually think is

[00:23:07] Justin Grammens: awesome. Yes. Yeah. Yeah, no, I totally agree. I mean, so we started Applied AI in 2019 and. You know, in 2020, March of 2020 ran into the same sort of problem that everyone did.

You couldn't do in person events and we were always in person before that. And then decided to just use Zoom. And so our model is a monthly meetup. We always do the first Thursday of every month, but the byproduct of the Zoom thing was that now I have speakers from all over the world, right? That are able to now present to the team.

And we just started going back with a hybrid model in June. So sometimes we have in person, sometimes we have virtual. I'm actually want to do the speakers virtual, and then the people are still here, right? So. You can attend and network with everybody else, but then the presenter is just on the screen, right?

And so, yeah, it's created this, this world because I've run a number of different meetup groups and technology groups over the past 15 years or so, but it's created this world that I never really thought that I would like, but it's kind of, can you kind of sort of fuse the best of both worlds? Because at the end of the day, it's really around about knowledge, right?

It's really about sharing and being able to get the most amount of impact that you can to your. to the people joining. And oftentimes that's not, they're just not physically within the same city.

[00:24:17] Rehgan Avon: A hundred percent. Yeah. I think it has really, really great benefits to it. You know, I don't think we would have been as aggressive on our virtual membership, like objective, like we never had chapters.

So we were always like, you know, big event and then see you next year. Yeah. Right. We launched the membership just to say like, okay, let's stay connected. And then the mentorship program came out of that and. Virtually, yes, it is hard, it can be, but it's, it's connects you with different types of mentors that you probably would have never gotten locally, you know?

Yeah,

[00:24:48] Justin Grammens: good. So we have, we have like liner notes and stuff like that. So we'll put links off to all the organizations, your company and the Women in Analytics organization as well. But yeah, are you always looking for mentors? Are you looking for people that want to volunteer?

[00:25:01] Rehgan Avon: Always looking for mentors, always looking for volunteers.

We just, I think we're about to open up Call for Speakers for next year as well. We have a very, rigorous process of speaker selection. And so we love getting as many applicants as possible. We do kind of a double review process with a content council that we select. So we select, you know, a number of speakers from the previous year in different areas of expertise, and then we kind of double review all the submissions.

So we're very, very intentional. about selecting speakers. So we'll definitely send you the link.

[00:25:32] Justin Grammens: Cool. That sounds great. You know, one of the questions that I do like to ask people on here is like, how would you define AI? You know, somebody says, what do you do, Regan, for your job? And you're like, I work in AI.

Well, what does that mean? Can you unpack that a little bit? And there's no right or wrong answer, of course, you know, but I'm just always curious to know what people, how they maybe describe artificial

[00:25:50] Rehgan Avon: intelligence. Yeah, this is such a fun question. And I think it's so fun because if you pulled like 25 people, In the space of a hundred people, they would all have a different answer, which I think is so interesting.

And so I started thinking about this question very deeply because people are very opinionated. Like they either sit in the camp of that typical icon that has the big AI bubble and machine learnings under that and deep learnings under that, that camp. Or they sit in the other camp, which is like, Oh, machine learning is not AI.

It's actually the opposite, you know? And so I think it's so interesting. You know, some people will describe this kind of goalpost that keeps moving because they talk about like the Turing test and what we perceive to be, you know, replicated intelligence via machine versus, you know, general intelligence.

And so I like to take the philosophical debate of what is intelligence. And, and then make it artificial. Okay. Yeah,

[00:26:46] Justin Grammens: sure, sure. Having computers do it.

[00:26:48] Rehgan Avon: You know, what, what do we consider to be intelligent? There's all the way at the plant level. Plants are very intelligent. And then you've got, like, the Elon Musk's of the world.

Right? So, there's this spectrum of intelligence and I think artificial intelligence very much falls in that same spectrum. There's a lot of things that we can do with machines that are intelligent. And some of that falls into these fun deep learning techniques and some of it doesn't. So I tend to take the more broad approach of artificial intelligence.

And I usually use a very simple example if folks are not from this field, which is like as a child, you learned to cross the street or not to cross the street by picking up on some very key signals. So your brain creates a model of. Do I cross the street or not? And it evaluates those key signals that you've trained it on, which is, you know, look both ways and see if there's a car coming and look at your parent maybe and see if they're starting to walk.

And, you know, there's all of these things that we capture and model in our heads. And we're just doing that with computers. And so, yeah, I, I find it to be such a fun, nuanced question.

[00:27:48] Justin Grammens: Yeah, for sure. As I said, this can be so broad with regards to applications and some people want to get, you know, down into the details or out, you know, sort of how it's done.

But, you know, at the end of the day, we are trying to program computers or some sort of system, I guess, to do things intelligently. And that's probably where the beauty comes in. What do you define as intelligence? A lot of ways. The other thing that I like to ask people too is, is like, so if I'm graduating from college, for example, like where do you suggest I go?

Like, how would I get into this field? You know, maybe for as a younger person sort of coming into this area.

[00:28:21] Rehgan Avon: Yeah, I get that question a lot. Also, a lot of individuals who want to kind of lean into this space and I've actually had many conversations with folks that are earlier in their career talking about this and the different options, like maybe they don't want to be super technical.

It used to be. Go get a math degree or go get a PhD in stats and like become a data scientist. And that's the only way when I started in the industry, that was the narratives. You had to have this, you know, everybody saw that graphic. What was like computer science and stats and math and business and tele, right, business domain expertise.

Everybody saw it all the time. I think we're in a really fun time for AI where there's a lot of different roles. You know, you've got people who are thinking about the ethical considerations. So you've got ethicists and you've got folks that are looking at the societal implications of it. You've got lawyers, AI lawyers, you know, who are thinking about interpretability of law and that impact on the AI.

Ecosystem and even down to like IP ownership now, right? And so you've got like kind of the legal aspect of that. And then you've got, of course, all of your technical roles. And then there's also what I'm kind of referring to as product management for AI or product ownership, which is really this idea of System design, you know, thinking about all of the different stakeholders and implications of these systems, coordinating the technical teams to build and evaluate different solutions and looking at feasibility.

Like, you don't have to be super technical to do a role like that. Some people might argue with me on that, but, you know, somewhat technical that maybe not hands on keyboard type of person. And so whenever I talk to someone, I usually start with the landscape. Like there are all of these different types of roles that you can take on.

And then you can back into like what school to go to, what degree to get, whether or not you need to do that. But that's kind of always been my approach.

[00:30:13] Justin Grammens: Yeah, for sure. I mean, there's these new titles. Like one of them is called prompt engineering, right? So that is somebody who is more a linguistics person or an English major than they are a hardcore.

computer programmer, software engineer, right? Yeah.

[00:30:28] Rehgan Avon: One of the most interesting projects that I was a part of in NLP, he was a linguist, you know, PhD who learned how to program. So, you know, you've got people who really, really want to understand kind of the domain of the area they want to be in. Which is another point that I always call out is, do you want to be a generalist or do you, do you care about a domain very specifically?

And like, do you want to dive into that? Yeah,

[00:30:55] Justin Grammens: yeah, yeah, yeah, for sure. And there's, again, there's so many different areas of AI that I don't even really touch, right? I mean, I don't call myself a cybersecurity AI expert in a lot of ways. When I go into companies, I will bring somebody else with me that maybe knows that side of the business.

But yeah, you can definitely focus. And, you know, this product management thing is really interesting. We, we did an event last week that was on product management and AI through the Applied AI group. And we had like 50 people sign up. And I was just super surprised that there's a lot going on in the space of product managers, how they can think about how to use this new.

Technology, build better products, to assemble a team to build better products, to work through this, this entire life cycle of product development, and yeah, the guy who ran it, somebody I've worked with for years, he would not call himself technical at all, with regards to being able to, you know, do the algorithms, he obviously has worked in technology, but Yeah, he just has a passion for this and there was a lot of people that showed up that were very interested in new tools, new techniques, and how they can bring these, this capability, I think, into the products that they're

[00:31:56] Rehgan Avon: building.

Yeah, a hundred percent. I'm actually speaking at a product leadership summit in October about this. You know, I, I'm biased. I, I have product background as well as technical kind of like hands on keyboard background, but I'm very passionate about this because the problem has always been system design, in my opinion.

It's always been the fact that we've never thought about the user very deeply. So I've actually seen a lot of UX researchers and designers becoming very knowledgeable about this because there is a paradigm shift happening in terms of, you know, user experience with products and incorporating AI into that.

Like, even down to, like, feedback loops. How do you get someone to participate in the feedback loop without knowing they're participating in a feedback loop, you know, that's the best kind of design that you can think of. And that provides a better performance for the model and so on. So I think that kind of stuff is very

[00:32:49] Justin Grammens: fascinating.

Yeah. Yeah, for sure. For sure. Well, and even around persona generation and sort of questions and answers. And so, I mean, imagine personas. around people that are using your products, and they're all essentially AI models, right? That you could then ask questions to. So, you know, it's, you can always get humans to ask questions, but we were going down this path, I was just sort of thinking back during the conversation we had for like two hours, one of these things, somebody was talking about prompting, and setting up a number of different personas, and testing them against your product, by asking them questions around features, and feature definition.

And, again, it was just basically, you know, chat GPT is what it was under the covers, but. It provides them with all that information that they can interact. And certainly they could scour the internet and search for all this stuff, but it's just all packaged up in a, in a language model. Yeah, it's so

[00:33:38] Rehgan Avon: fascinating.

I used to do like simulation engineering work for like in automotives years ago, probably over 10 years ago at this point. And, you know, that was us basically simulating, you know, a manufacturing line and then moving things around and doing design experiments and changing things. But we took data, we modeled it, and we created a simulation of that manufacturing line.

And there's, and now that we can get really good. With these language models, we can start to do simulations of people like you're mentioning. Like you can start to simulate these kind of user experience feedback loops and user personas and things like that, which is really cool.

[00:34:15] Justin Grammens: Yeah, this leads me to maybe one of the final questions that we'll have.

But just how do you think this is going to change the future of work for people? So obviously some people are saying, well, you know, now we don't need product managers anymore. Somebody else can basically write all this stuff out or have an AI write it out. How do you view AI sort of changing our careers and what we do as humans, I think, in the next, you know, five to ten years or so?

Yeah,

[00:34:39] Rehgan Avon: this is, um, a very hard question to answer, but of course, you know, I'll answer it with the knowledge that I have today. Yeah. I think in the short term, We're going to see a lot more interesting participation in individuals who want to be a part of kind of the bigger strategic elements of their job.

So, you know, you've got somebody in supply chain who is on the operation side and wants to be able to participate in AI design and development. And their job becomes more kind of creative and interesting as they think about AI implemented into their Their operational world is, you know, in their job, and then you've got individuals that maybe aren't as excited to do that, and obviously folks see this as an opportunity to drive more efficiencies inside of their organization, cost cutting exercises, things like that.

I didn't, there will be a decent amount of that that happens, but I also believe that. There will be more kind of interesting job functions surrounding those automations and we will need people to guide these systems for a while, for a long time. And so I think in the short, near short term, you know, that's going to continue to happen and people don't like change.

And so it's going to be slow. I think it's always be slower than what folks anticipate, but I have no idea after that.

[00:36:02] Justin Grammens: Yeah, I think the human creativity always seems to kind of win in the end, and I take a positive, you know, look that it will be an augmenting technology that will, you know, help us do our jobs better.

Yes, and maybe that's what you're touching on a little bit. Some people won't want to bring this into their job, and they're probably the ones that are going to be losing their job. But those that can leverage it and bring it into their job and work with it to the best of their ability and suss out these efficiencies will always be able to be creative with.

And AI is not the last technology, right? I mean, it's just, who would have thought of the internet? And who would have thought of cell phones? And who would have thought of all the stuff that's been going on over the past, you know, 50 years or so, right? Or even more! Right? Since the Industrial Revolution.

So, there's going to be the next thing that's going to be sort of earth shattering or game changing, and I just, I feel like humans have always been able to.

[00:36:54] Rehgan Avon: Yeah, I agree. I think people do try to parallelize this paradigm shift to others, and I think that's a good way of thinking about it because that's kind of our reference points.

You know, I do think this is probably somewhat unique, just in the sense that it's a magnitude of productivity that we can unlock, maybe greater than what we've seen before. So that's why I'm always so hesitant to say what it's going to look like on the other side. I know a lot of people have. Very, very strong opinions on that question too.

Yeah, but I, I generally agree with you, I take, I tend to take the more positive approach to it. Yeah.

[00:37:28] Justin Grammens: I will agree with you that there is something different here. There's a author, you've all know, a Harari, I think he wrote a book called Sapiens, which is a really, really good book, but he was interviewed recently and I'll make sure to put this in the notes.

He was interviewed by Lex Friedman, who does a really interesting podcast. But Harari had, had, he's just, he's phenomenal. He's somebody who's deeply thought about all sorts of stuff. Not only the history of human civilization, but also how we're headed in the future. But I'm going to completely butcher what he was getting at.

In general, he was saying that AI is a little bit different because if you think about, you know, everything that's been invented up to now, like if it's a knife, or it's the wheel, or it's a phone. Or something like that. It's always been a tool that we could basically use. This thing doesn't do anything without actually human intervention and the human actually, like the gun, for example, right?

There's all this new technology that's been coming out, even the printing press, right? This is different in so much as that it's generative, right? It's something that can actually do and provide intelligence better than what a human can. And he was saying that nowhere in history until this has that really ever, you know, occurred.

These have always been dumb technologies in some way, and then it's needed a human to sort of like supercharge it. And a lot of ways with AI, it can basically take control. You know, we maybe have drones, smart drones that are already flying. They can actually do things in the physical world even better than what humans can do.

So that's what I think has got people a little bit scared and worried about. Yeah,

[00:38:52] Rehgan Avon: and I actually tend to err on the side that that's very real. I don't think that's here yet, but I, but I agree with that. Yeah,

[00:38:59] Justin Grammens: yeah. So it is a different paradigm shift for sure. Well, Reagan, this has been a great conversation.

Is there anything else that you want to maybe talk about, I guess, or questions I should have asked or whatever that you'd want to share with our

[00:39:09] Rehgan Avon: listeners? Yeah, I think, you know, you had mentioned, I always talk about resources and people to follow because I'm big on that, too. So there's always a lot of women that I think are incredible to follow.

There's a couple of ethicists like Renee Cummings. There's some other women kind of at the forefront of this, like Chip Nguyen on the machine learning, like Emma Love's side. Cassie Kozyrkov is obviously amazing, and she's great to follow. On the decision science side, Noel Silver. Also, unbelievable. She's at Accenture and she has a deep background in this space and was one of the original contributors to Alexa.

So I always like to call out some other incredible women who are doing some really incredible things in this space that people can follow for more information and content.

[00:39:49] Justin Grammens: That is awesome. And I should get these names, or at least introductions, to have them on my podcast in the future. For sure. Um, this podcast has been great.

I mean, you know, this will be episode 73 or 74 or something like that once this goes out. But it's been really fun because I met you through somebody who I met through who somebody I met through somebody I met through, right? So the branch has just been super awesome. It's been super fun. And every time I have somebody on the podcast, I usually try and ask them, Hey, can you introduce me to somebody else?

So. I've got about three or four different names here that I'll, I'll, I'll want you to connect me with if I could. Amazing. Yeah.

[00:40:22] Rehgan Avon: Happy to.

[00:40:23] Justin Grammens: Awesome. Well, great, Regan. I appreciate your time and all the insights that you've given and best of luck with Align AI. Sounds like a, you know, it's a two person organization right now, but you're probably looking to grow in the future.

[00:40:33] Rehgan Avon: Yeah, we've actually grown quite a bit recently. So, and we're about to grow quite a bit more too. So yeah, we've got maybe like. 10 folks working on our, our product right now. And yeah, about to hire again. So head over to our website, definitely, and check

[00:40:49] Justin Grammens: us out. Yeah, yeah, for sure. You mentioned you and your co founder.

So I wasn't how big the company was, but congratulations. That's awesome. You have 10, 10 people and yeah, taken off in this new fun and amazing world of artificial intelligence. So. Thanks again for your time and look forward to keeping in touch.

[00:41:02] Rehgan Avon: Yeah, it was a pleasure. Thanks for having me on.

[00:41:06] AI Voice: You've listened to another episode of the Conversations on Applied AI podcast.

We hope you are eager to learn more about applying artificial intelligence and deep learning within your organization. You can visit us at AppliedAI. mn to keep up to date on our events and connect with our amazing community. Please don't hesitate to reach out to Justin at AppliedAI. mn if you are interested in participating in a future episode.

Thank you for listening.