EPISODE 1626 [INTRODUCTION] [0:00:00] ANNOUNCER: This episode of Software Engineering Daily is part of our on-site coverage of AWS Reinvent 2023, which took place from November 27th through December 1st in Las Vegas.  In today's interview, host, Jordi Mon Company, speaks with Kevin Kelly who is the Director of the AWS Cloud Institute. This episode of Software Engineering Daily is hosted by Jordi Mon Companys. Check the show notes for more information on Jordi's work and where to find him.  [INTERVIEW] [0:00:37] JMC: Welcome, Kevin, to Software Engineering Daily.  [0:00:40] TR: Thanks, Jordi. Thanks for having me.  [0:00:42] JMC: Would you mind introducing yourself?  [0:00:44] TR: Absolutely. My name is Kevin Kelly. I've been with AWS for 14 years and 10 months now. I've really seen the cloud evolve from a small number of services to lots and lots of services and lots of GenAI capability.  [0:00:57] JMC: We are at Reinvent 2023 in Las Vegas, Nevada, the US. How many Reinvents have you attended already?  [0:01:06] TR: I've actually attended all of them but one. [0:01:07] JMC: Oh, wow.  [0:01:08] TR: Including the first one, which was really something.  [0:01:10] JMC: Indeed. How has the evolution been? Because I'm not sure if it's record-breaking. But Adam, the CEO of AWS said that we have around 60,000 people here. And it feels crowded in a good way. There's such a vibe here. How's the evolution of Reinvent gone through your eyes? In front of your eyes?  [0:01:29] TR: Yeah. It started as a very small event with a few conference rooms in a single hotel. And now we're taking up lots of hotels up and down the strip. There are lots of people. I can't believe the escalators don't break.  [0:01:39] JMC: Yeah. AWS just basically takes over Las Vegas.  [0:01:43] TR: Yeah. And with all of our partners here and the customers that come to the event, it's just unbelievable how many people are here and how many different things they're doing with the cloud and talking about in all of the sessions.  [0:01:53] JMC: What's your charter at AWS right now?  [0:01:56] TR: Yeah. Right now, I lead the AWS Cloud Institute, which is an exciting program that we launched on 10 October. And our first classes begin on January 8th. I love to talk more about that later. [0:02:06] JMC: Exactly. I get requests for consultations about the software engineering space, DevOps space, many things. And one of the data points that most of the sort of like investors or whom ever's paying me for consultations or anyone in my position asked me about the supply of software engineers. They care about cloud-native, or mainframe, or the different segmentations. And that is always a difficult calculation that I have to do. But the context is absolutely clear right now.  The supply of whether sofware engineers, operations people, cloud engineers is not following the pace that specifically AWS is demanding. Because you guys are moving at the speed of light. In that context, I believe that many of your educational offerings and just institutes, and they are not necessarily at all commercial offerings, have been announced today and are being really used out there. Could you give us a lay of the land of what AWS right now offers in terms of educational purposes? [0:03:05] TR: Yeah. We have quite a few training programs. And we're really trying to meet learners where they are. We have programs like AWS GetIT and AWS Educate which are targeting young learners. As early as primary school and secondary school. And we're providing experiences for them where they can use the AWS console and actually do labs in addition to taking digital e-learning courses on a lot of different topics.  We have programs that are targeting higher-ed institutions. We have a program called AWS Academy where we give curriculum and lab access to educators to use in the classroom and not just in computer science degrees or business IT degrees, but all kinds of degrees. [0:03:46] JMC: Oh, okay. For different domains. Other than the technical computer sciencey ones.  [0:03:50] TR: Yeah. [0:03:51] JMC: Oh, wow.  [0:03:52] TR: And as the cloud becomes more pervasive and GenAI becomes more pervasive, everyone's got to begin learning those skills. And if you have a cellphone, you're using the cloud today whether you realize it or not. We have programs that are available for our corporate learners who are getting paid to use AWS building all kind of amazing things.  We actually have AWS Skill Builder, which is a platform that has hundreds of free digital learning courses, and content and modules supplemented with labs that our learners can use and learn on. We have programs like AWS Restart, which is a 9-week boot-camp-like program that culminates and our learners earning the Cloud Practitioner Certification and taking some of those entry-level less technical kind of positions in the cloud. Ideally, ramping up to some of the more technical roles.  And then, finally, we have the AWS Cloud Institute, which is the program that I'm involved with, which is as little as a year-long, 12-course program that allows our learners to earn the Cloud Practitioner Certification in the first quarter and the Developer Associate Certification by the end of the program. And we make no prerequisite assumptions about what they know. It's really for anyone that has the grit that wants to go through it.  We teach Python as the coding language. And we start by explaining application program interface, or API systems, distributed systems and virtualization, which was sort of the perfect storm that the cloud was born out of. And we build on it from there and start talking about fungible compute, storage, network, database. And then we layer it with topics such as cost optimization and performance. What if I want sustainable applications?  And we really lead them through a series of complex topics supplemented with over 100 labs. They build two Capstone projects as part of it all included in the program. And we give them 50 gigs of AWS CodeCommit to save everything they've built so that they have a portfolio of assets when they want to go get a job. [0:05:52] JMC: And you have set yourself - because, again, this is not only to sort of like supplement the lacking supply of talent for the cloud, but it's also that you have set yourselves a very challenging goal to train a lot of people with this set of programs, right?  [0:06:10] TR: We have indeed. In fact, Reinvent in 2020, we announced we wanted to train 29 million people by 2025 for free. And I'm happy to say that we're over 21 million towards that goal as we sit here today.  [0:06:22] JMC: Two years to go still.  [0:06:23] TR: Yeah. With some time left. [0:06:25] JMC: Yeah. And then in November, we announced our AI Ready Commitment, where we really want to help young Learners begin learning about the cloud. And we want to train 2 million of them by 2025. And we have some fun programs. We have some new and free AI and generative AI courses, but we also have a scholarship for 50,000 high school or secondary learners and university learners in conjunction with our training partner Udacity. And then we have an hour of code dance party, which is an event that helps teach GenAI. And we really want to meet those young learners where they are with fresh topics that are of interest to them. [0:07:01] JMC: Obviously, many newcomers to these programs or specifically the GenAI ones are excited about the hype there and very reasonably so. But I mean, the CEO of AWS said it on Tuesday at his keynote. I think Werner this morning said it. We are the beginning of this revolution. In your opinion, in your experience rather, what are the biggest gaps in terms of skill building with regards to GenAI? What is right now what more people are demanding for in that concrete space?  [0:07:32] TR: Yeah. People are definitely looking for cloud builders that can help build and maintain large language models that are driving a lot of the GenAI experiences. But even things like AWS CodeWhisper is something we address in the AWS Cloud Institute. And it's a bit like long division. We teach our learners Python from the ground up before we introduce CodeWhisper. Kind of like when you got to use a calculator for long division, you're like, "Why did I ever do it the other way?"  And then we teach them not only the mechanics of prompt writing and how to begin using CodeWhisper in the context of being a cloud builder as an assistant and an aid. But we talked to them about where do my prompts go? The code that's generated, where is it coming from? What license agreement is associated with that code? And we really get them to think a little bit more responsibly about the context of these kind of GenAI tools that they're using to become a cloud builder. [0:08:24] JMC: For reference, I interviewed Deepak Singh from AWS a month ago. No. Probably a bit more. I can't remember. But one of the things that we spoke most about was CodeWhisper. And I would argue, things move really fast. But I would argue that, certainly, at that point, one of the key differentiators of CodeWhisper was that it was able to reference. I mean, there's no direct link from training data and the output of LLM. Whether it's the one powering CodeWhisper or any other. But it's able to suggest a link between the training data. In this case, open source, and signify and remark what license is there. Very good open-source citizenship there. But yeah, it's one of these amazing co-pilot, pair-programming tools that are there. I'm happy that you teach them the hard way first and then show them that AI is there to help them.  What else? Any other - I mean, Ops is not going away. I mean, the wide variety of skills that anyone playing around with a myriad of AWS Services require is just humongous. But any other? Do you see a huge demand for training and skill-building?  [0:09:32] TR: Yeah. The basics of the cloud are still required. In almost all of our training modalities and courses, we talk about what is fungible computer. What does that really mean? Disposable and reconstitutable compute. In the old days, you had servers on-prem in a really cold room and you named them Simpsons characters. And they formed unnatural attachments to hardware. And that's gone.  And getting learners to think creatively about compute is something I can dismiss and bring back on demand in a scalable way. And that leads to the topic of elasticity. And how can I use elasticity to embrace increases in workload and then dismiss it when I don't need it and don't want to use it? And how can I build applications? And what does that mean for sustainability? If I can give up that compute, make it available to somebody else so that I'm not wasting it and the electricity driving it and powering it.  We really try to tie a lot of those concepts together. Those are the basics of compute, storage, network, database. And then when you layer on top of that things like sustainability or the choices I can make around performance versus cost optimization, it can get really complex really quickly. And we're trying to help our learners understand how to navigate those topics and understand the tradeoffs that they need to make when they're building solutions for the cloud.  [0:10:50] JMC: I was going to ask, you've mentioned a minute ago two certifications. The ones that one would get if that person passes the Cloud Institute Program. Correct?  [0:11:00] TR: Yeah.  [0:11:00] JMC: But those are two of more, right? There are more certifications that AWS offers, right?  [0:11:05] TR: Yeah. We actually have over 10 certifications on a bunch of different topic areas. As part of the AWS Cloud Institute in the first quarter, they get the AWS Cloud Practitioner Certification, which is a general knowledge certification.  [0:11:16] JMC: Entry-level.  [0:11:17] TR: Yeah. About the cloud and AWS. And then towards the end of the program, they earn the AWS Developer Associate Certification. And our certifications really let our learners demonstrate that they understand the cloud and the services that AWS provides for the cloud in a deep and comprehensive way. [0:11:35] JMC: Okay. Any other certifications that you want to highlight that are quite in demand lately or not?  [0:11:42] TR: Well, our Machine Learning Certification, as you might imagine. Our Security and Networking Certifications are always in demand because those are topics that people really need to understand deeply. And the certifications give the professionals that have that knowledge the ability to demonstrate to an employer or a customer that they really have that skill set.  [0:12:01] JMC: I'm fairly sure that, throughout your career, you've found people that have pivoted in their professional careers, right? I mean, we haven't talked about this before. But have you found any really surprising and breathtaking sort of like career pivots that now have provided through the educational programs that AWS provides a new life for someone that was bored of his old job or actually had hit a limit and could not progress anymore or promote? Have you come across any of these stories or any particular people that you found that touched your heart?  [0:12:33] TR: Yeah. Well, there are two that come to mind really quickly associated with our AWS Restart Program. One is a hospitality worker that, when COVID happened, they were out of a job. And the other one was a musician. A professional musician that had served in the military as a musician and was transitioning to the civilian workforce and used our training program to help bridge that gap into the civilian workforce and learn about the cloud. Got an entry-level position in the cloud and is still there today. [0:13:03] JMC: Oh, nice.  [0:13:04] TR: I love those stories where these are mid-career changers that are changing their situation and using these kinds of programs. But we have lots of stories people using our AWS Educate Program and picking up badges and taking specific courses and then using that to lever into a new job. [0:13:22] JMC: That's the beauty of all of this, in my view at least. Musician jobs are not going to go away. Hospitality jobs are not going to go away. But the truth of the matter is that cloud jobs are going to be everywhere. The fact that you don't need to go to uni, or you can go to UNI but complement your training there, your education there with the stuff that AWS provides is just fantastic. It allows you to pivot to a well-paying, stable, with future reliance job like almost anything related to the cloud. I'm quite grateful.  What other announcements has the AWS made? In these regards, AWS. you've mentioned the commitment to training millions of people by 2025. That was years ago. But what else have you announced this specific year?  [0:14:08] KK: Yeah. This year, we did announce our AI Ready. And it's a commitment to train those 2 million young learners by 2025. And they'll be net new to the 29 million that were already training. And then we made our commitment with the AWS Cloud Institute that we announced this year. It's interesting when you bring up higher-ed. We have a program, AWS Academy, where we provide curriculum and labs to educators in colleges and universities to teach those learners. And for us, it's an and. Not an or. And we have programs like Restart, which is a 9-week boot-camp-like program that gets you to cloud practitioner.  And we have a program in the middle, Cloud Institute, where we want to pick up those two certifications. And all of this being said, part of our training modalities are gamified. You might have seen our Cloud Quest, interactive, gamified learning environment where learners can walk around in a virtual reality environment and learn things and interact with the environment while they're learning. We use that in some of our other programs like AWS Educate and AWS Cloud Institute.  We have an AWS Jam platform where learners can get together and have a bit of a challenge and compete as teams and learn along the way. And so, we're really trying with all of these programs to meet every learner where they are that has an interest and wants to get into the cloud. [0:15:23] JMC: It reminds me of the vibe that PartyRock gave me. I interviewed a week ago Mike Miller. One of the AWS workers involved in that. And it's not an educational program. Although, it is a product per se. It is an application or an application builder. But it has a fun experience and an educational purpose. Because as a user of PartyRock - and PartyRock, I welcome everyone to just go to the website. It's a very easy way to build. Quite simple, but fun applications. And modify them. And understand the underlying prompt, the underlying fine-tuning. Not complete. There's not a huge variety of features, but there's some. Like, Temperature, for example, that increases the randomness of the output of the LLM and so forth.  I really like how in tune that product is with the spirit that many of the programs you guys have put out. I wonder if anytime you would be able to put PartyRock inside any of these programs that you've mentioned. Maybe it will become a tool for education. [0:16:24] KK: Yeah. I don't see why we wouldn't.  [0:16:26] JMC: Have you played around with PartyRock yourself?  [0:16:27] KK: I have not. But we've done that with AWS Jam, and Cloud Quest and our other gamified offerings. We really want to put as many of those into as many training programs as we can. [0:16:37] JMC: What is next then? What is going to happen in the future for AWS educational programs? The goal is quite ambitious. But yeah, how are you going to achieve that?  [0:16:49] KK: I think we want to help our learners stay current with the state of the technology. And GenAI hit us like a storm a year or so ago. And we're responding with training programs for young learners, for mid-career change learners and corporate experienced learners in the active IT workforce today. Meet them where they are with training programs that get them educated on and trained on GenAI so they can build amazing things leveraging that new technology. And I think a year, two years from now, whatever the next big wave of technology is, we're going to want to follow that similar pattern. [0:17:22] JMC: Yeah. Do you have any forecast of what's coming? I see that agents might be something really popular. But I don't know. What's your opinion about what might come next?  [0:17:32] KK: I think we're not even close. And I think GenAI is actually going to help in the AR/VR learning space and getting more learners more immersed in those virtual reality, augmented reality environments and having them doing amazing things without doing anything in a physical environment yet. Cutting their teeth there and then transitioning that into a physical environment. I think we're just at the beginning of that.  [0:17:57] JMC: Is that you sort of like gauging the interest of the youngest of those that have played around with the programs that you deliver for them? Or is it more an opinion about like an informed opinion from a professional with a long career about where things are going? All I'm interested is in knowing how are the young cadets of the world using the educational programs that you guys put out there? For what purposes? What are they looking to build? AR applications? Virtual reality ones? Or maybe other things? What is your opinion on that?  [0:18:27] KK: Yeah. I think the appetite especially for younger learners, but older learners as well, we're all getting a lot more comfortable with a lot more modalities. Visual, audio, combination of those and then using those to learn new things. 10 years ago, how many of us were watching YouTube videos when whenever we need to do anything? Now I don't go do anything without looking to see if there's a video first to show me how. And I'm learning how to do things I didn't know how to do before. And I think as all of the learners get more comfortable with those kind of modalities, we want to meet them with gamified AR/VR experiences that help follow those modalities?  And for visual learners, experiential learners, kinesthetic learners, these new modalities are opening up education paths that might not have been there if they don't like reading 5 hours a day and were a really good standardized test taker. And we really want to reach those learners and bring them into the cloud IT workforce. [0:19:22] JMC: As a teacher, as a person in charge of educational programs, how do you approach this? It seems clear that - and gamification is a fundamental part of the way you build the content. But it's also true that certifications, for example, the ones that you mentioned, are quite difficult. There is a degree of knowledge that you need to reach. And that is through learning. How do you strike the right balance? How do you approach the creation of educational content for any of these program certifications and so forth?  [0:19:50] KK: I think it's about getting learners to successful outcomes as quickly as possible. Gamification and those kind of modalities can be a hook. And they're fun and they can get you started. But nothing beats the rush of learning something and building it and having it work. And the cloud's a great environment to do that.  And so, we try to get learners there as quickly as possible. And then we want to give them confidence and swagger. Build the next thing that's a little more complicated. Now take that and build something different and hook them together. And as their confidence increases, that's really where I think the learning takes place. And then, the certification exams become a foregone conclusion. Because I deeply actually understand the services, the capabilities. How they hook together in different combinations? And performance, and cost optimization and everything I need to wrap around it.  And really, for our learners that go through those kind of training experiences, they tell us passing the exam was the easy part. Because they get confident building and using the systems. AR/VR, gamification, as many modalities as we have. Digitally learning, instructor-led, synchronous, recorded, playback, caption, translated. We really want to give them as many learning options as possible so that they get that confidence.  [0:21:05] JMC: Yeah. And as you said, most of the outcomes in the shape of the software that this people have built will then eventually help them have a portfolio, which is one of the biggest reference you can make when applying to a job. It's like, "You know what? I not only say that I did this, but here you are. This is the application." The fact that this can be stored, and maintained and kept in AWS, it's just absolutely brilliant.  [0:21:30] KK: And I think employers are embracing looking at assets and portfolios of things you've done. In some industries, it's always - if you were an artist, of course, you always brought what you've done to show this is what I do. It hasn't always been in the technical fields. Now they can bring that code when we enable that in some of our programs so that they can keep the assets that they build. And nothing can be more convincing and powerful in an interview than you walking through something you built as the expert. [0:21:59] JMC: Yeah. Any chance that any other programming languages will be incorporated? I mean, Python seems like a very logical choice. It's one of the most understandable syntax out there. But have you considered any other maybe low-level languages? Would it make sense at all or not?  [0:22:15] KK: Well, I think the trends would have us moving towards more and more higher-level programming languages. And prompt writing is becoming a topic today. I think we'll continue to lean in there. And we'll follow the market and the trend and we'll pick the languages that our customers and partners tell us they're looking to hire people with. [0:22:32] JMC: Okay. What else? I've seen there's a lot of builders out here. In fact, I think - well, no. Werner this morning in his keynote called out the champions, if I'm not wrong. But anyway, I've seen plenty of builders here. Any application? Any piece of software that has caught your attention lately that program participants have built? Would you highlight anything that has been fun and has caught your attention?  [0:22:57] KK: There's one program we had that contained capstones. We had some young learners who were introduced to AWS as part of this program. They had an interest in music. And so, they built a system that listened to the duration and frequency of music and they created the sheet music from music that they're listening to.  [0:23:16] JMC: Oh, okay. The other way around. And instead of composing and then playing and performing, this thing was listening to the music being played. And back engineer -  [0:23:26] KK: Yeah. Sort of reverse engineer the sheet music. And they did it. It was a bit of a passion project because one of them was a musician. But he said I'm not really good at listening to something and then being able to play it. But I can read the notes. And so, they kind of built it to help themselves. And for me, when people have passions like that around specific tech, that's where the innovation happens.  [0:23:48] JMC: Yeah. Do you remember what was the specific of the stack they've built? I mean, I know you were not about the builder. But how did that came to fruition? [0:23:55] KK: Yeah, it was part of a Capstone project. And so, they cobbled together 5 to 10 AWS services to actually accomplish that.  [0:24:02] JMC: Okay. Wow. Fantastic. Incredible. Yes, I think that, other than that, we've covered everything that you guys have announced. Again, the goals are incredibly ambitious. And I'm really happy that AWS is doing this to complement universities, high schools, private schools, everything. Because, I mean, the jobs in this market is just - the number of people that AWS will hire in the future is just ridiculously high. The jobs are going to be in demand for sure. Yes. Other than that, I have no more - nothing else to cover. Did we miss anything that you consider worthwhile touching upon?  [0:24:36] KK: Other than we want as many people who have the desire and want to get involved in the IT workforce, to take a look at our programs, other companies' programs. Just get here. Learn the cloud and join us. [0:24:50] JMC: Where can they find everything that we've talked about?  [0:24:53] KK: At awstraining.com, they can find all of our programs. [0:24:56] JMC: Okay. I know there are not only courses and certifications, but there's paths to lead anyone down a structured and formalized path, right? Which is really convenient. And many more things.  Well, Kevin, thanks so much for being with us. All the best with achieving those goals. Because it will require a lot of work. Although, you've completed almost all of the goal set for 2025 as you just explained. But it's incredible that you guys are putting this out. And really grateful.  [0:25:19] KK: Thanks for letting me talk about it, Jordi.  [END]