Panel: How Platforms Can Save Junior Engineers (and Thus... Jennifer R, Leena M, Molly C & Paula K

This talk focuses on the crucial role of platform engineering in supporting junior developers within tech organizations. The speakers, including experienced engineers from various companies, share their insights on how platform engineering can facilitate the growth and efficiency of both junior and senior engineers. They discuss the impact of AI on the engineering landscape, revealing that junior developers often adopt AI tools more readily, as they help mitigate the lack of foundational experience. Throughout the session, the speakers emphasize the importance of creating an inclusive platform that not only aids in software delivery but also empowers junior developers by reducing cognitive load and providing critical resources. The conversation highlights the need to integrate junior engineers into teams to encourage creativity and innovation, ultimately ensuring the sustainability of the tech workforce.

Hello everyone and welcome. We certainly think this is an important topic. So I'm so happy to have you all here. So I like audience participation. So the lightest is a straw poll. And then we go up to if you want to ask questions at any point, especially if you're a junior engineer and want to share your experience, it's totally fine. There's a microphone right there in the

middle. You may quue up and I may or may not call on you at some point. But in the meantime, I want to take a straw poll. Raise your hand if you work for a company with a platform team, platform engineering team. >> Okay. Raise your hand if you work for junior engineers or work with, sorry, work with junior engineers. Important preposition. Raise your hand if you

are a junior engineer. Thank you for coming. This is it's more for the other people in the room because we want to help you to the junior engineers. And then who is worried about the future of junior engineers? Okay. So, we know where we stand. So, I'm really delighted to have these three amazing people, humans and tech friends on stage today. So I would like you to

both all three of you even to introduce yourself and then also reflect on where you got started as a junior engineer and where you are today in the platform space. Do you want to go first? >> Yeah, absolutely. Hello. Um I am Lena Munaram and I've had the pleasure of being a junior engineer multiple times. Um so once straight out of university and then again after having

switched um switched careers I suppose and industries and then coming back into tech. So um it was it's been an interesting experience all the way around. Um currently a senior platform engineer um for chain analysis who are the data platform for the blockchain. >> I'm Molly Clark. I work for EasyJet at the moment and um I started as a junior engineer basically in what you'd call a

platform engineering team today. We were lucky then there was a lot less technology to learn and we had small teams of about four to five people focusing on the automation and the builds of the different types of technologies like Seabore and Egate if anyone even knows what they are. Um, I then went through and I became a junior again when I wanted to retrain a little bit

as we went into the cloud. And now I look after a platform engineering team trying to build out a platform to support both juniors also new joiners and senior people across our company. >> Hi, I'm Paula Kennedy. I am the co-founder and COO of a company called Centaso. I also co-host the platform engineering day for those that were there on Monday for the collocated day. I started

my career 2000 so quite a while ago. Uh and I was a support analyst supporting IBM AS400 machines if anyone's got a clue what those are. So uh quite a long time ago. More recently I am as mentioned the founder of a company. So I feel like I'm a junior in in all things actually. Startup life means having to learn a whole bunch bunch of disciplines and

different things and really learning from scratch. So I feel like I'm junior all of the time now. >> And then there's the AI of it all where we are all more junior than we really think. So, I just want to share a couple things because I happen to be talking to Octopus's Octopus deploys Charlotte Fleming the other day and she had really dug into some of the

data because it's all open source with Stack Overflow from their state of the developer survey and they don't they can't possibly publish all of it. So, there was some of that she found interesting that I was like oo I wish someone had published that. So, good but they published the data. So, thank you for that. Developers with zero to two years of experience demonstrate the highest use

of AI across most tasks, particularly tasks that are foundational to skill acquisition. And this suggests Charlotte was talking about that junior developers readily adopt AI not because of resistance to traditional methods, but rather because they lack the baseline experience necessary to critically evaluate its outputs. So I think this is why companies think they can just not cut their cloud bill instead fire people. Uh because they and

they can kind of use AI as an excuse that but basically when we were discussing this in the start AI is basically another junior engineer. And then I just want to shout out the 0800 DevOps podcast. There was an episode on AI first development workflows and Michael Gates was talking about his team had worked hard to implement AI which eventually led to a 38% increase but not

straight away and then he found that his senior engineers were using AI like they were using claude and everybody got 500 prompts a month or whatever the deal is at that time and everyone got the same prompts and they were just not using it as much and the junior engineers were burning through it within two weeks. So he said new engineers don't have the context of years

of experience in multiple repositories and multiple hosting environments and all the nuances that come with development it. But that's always been true. That's most boot camps and certainly universities are basically building green field and then without that context you're at a big disadvantage which I think we know what creates that context platforms. So just because I really do want to stress that we want to hear from

the juniors in the room. I was talking to a friend who they retrained recently and just finished graduated from their apprenticeship in software engineering and I was like what do you think about platforms because we're really all in like we've on the stage we really like it. I was like, you know, we need to talk about it. And they said, it's definitely nice to have a centralized

web page that I could go to for questions, but honestly, it was way nicer actually meeting the team so I could have real people to reach out to. They said, and we found this with the door report and all that, they were asking better questions, they thought, but they missed that connection. So I think that's an important thing when we're looking at addressing platform as end all

beall. It's also we have to make sure we're asking the junior developers. So just one more thing in my intro. If you came here because you wanted to figure out how to use platform engineering to replace junior developers with and safely with AI, that is not the policy or the intent of this. So if you came for that, maybe we'll convince you otherwise. But we all very

firmly believe, hence the title, how platform engineering can save the junior developer role and save the tech industry is true. You can't have senior developers without junior developers. You're going to disconnect from your user base. You're going to break the world. So we definitely want junior developers. So that is something I think we all agree on. But so yeah to I guess we'll start with you Paula.

Uh what just to ground us in objectives what do you think is the objective of platform engineering and how can that apply to junior developers? >> So for me platform engineering at its most simple is about enabling organizations to go faster deliver software more effectively and more safely. It's designed to reduce cognitive load, not be a bottleneck, but be an enabler to all teams with an order.

That's what that's what platform engineering at the heart of is aimed for. And I think with junior developers, there's lots of ways that platforms, internal platforms can help. So in Centasa, we talk about kind of the four pillars. Safety, speed, efficiency, scalability. Those are the things we think of a good platform can do. And I lots of things you mentioned about having that central website for junior

devs to find things, right? Having your platform offer self-service capabilities means junior developers can get the things they need. It's true for everybody, junior developers and AI agents, but they can get the things they need when they need them. If you have guard rails built in, then they can deploy safely. Hopefully not screwing things up. I mean we've seen some horror stories in the news but if

the guardrails are built into the platform security governance that provides protections that helps junior developers and then kind of that uh scalability if you can make it possible for junior developers to contribute into the platform to be able to like be effective sooner that gives them that kind of motivation to keep learning to keep growing. It's um I mean as you mentioned we're all big fans of

platforms around here but it really can enable Julian >> Um yeah so that's pretty much it. Yeah. So platforms are here to help us um make it so everybody can deliver um better quality without breaking down bringing down production for example. And we were talking earlier it's about giving the developers the endorphin hit. They've actually managed to deliver something. They've managed to build something. it's gone green.

They're not constantly hit by barriers because they didn't know that they needed to get the scanning in place or they didn't know that they needed to get that review done or they didn't know that they needed to go and request the AWS account 10 months before they needed it. All of these kind of things that you know we're just trying to remove those blockers and make it

so you can focus on the work and not worry about all the things that other people have done before which are not value ad. But then platform engineers tend to be very senior roles. So is there a disconnect there or >> I mean you could say you you can have a mix of junior and senior um platform engineers as well. So I'm a big fan of bringing

in graduates and um new joiners because they are like really hungry to learn. Um, and of course there's always tasks that the seniors won't have time to do, and it's better to get the juniors to do them so that seniors can spend more time doing the, you know, the new sexy bring in an AI, right? And building those guardrails for that. Um, but yeah, it's you can

definitely have a mix in the platform team because again, you're going to need that. And it comes back to um I went to a talk yesterday and it was really enlightening for me because you know you end up building the platform for the noisiest and the experts right the people who moan about this and who want everything. So that's like the 20% whereas actually you want to

be thinking about what can you do to make it that everybody can deliver quicker and safer. Um so that's as said the juniors the people joining teams joining companies AI agents as well these days um you can't have it that you just have a couple of seniors and a million AI agents because then your seniors are drowning in cognitive load from that. So it's all about the

balance >> and I would I would actually argue as well that platform is a great place for juniors in that a key part of being a really good platform engineer is asking questions is talking to people and you know coming at it from a perspective where actually you don't know what the landscape looks like. Some of the questions that you ask may actually be the questions that

need asking and that will actually bring up um areas that platforms should focus on but perhaps the senior platform engineers are so or they're so used to the context that they don't see it anymore. They don't see the gap. >> Yeah. >> And they may also be kind of indoctrinated or somehow they can't see. I mean one of the things we try to do internally is when

a new joiner comes in we ask them to be like what's one thing you would improve because the newer people coming in know the least but also are the least kind of thinking of that mindset of it's just this is how it's always been and so the newer people who've got new ideas new perspective fresh eyes can come in and be like why is it like that

and then we try to encourage people to say change one thing you've done this on boarding change something you've done this new new kind of release change something constantly trying to bring in new ideas >> and you don't have that well this is the way it's always been >> because it's you know it hasn't if you're new to the team or new to the >> and I

guess it's hard as a platform engineer also to reach out to the 80% the and especially the junior engineers because they're less likely to like we said it's it's the flip side the senior engineers are going to be more vocal about and the more advanced about what they want but also especially in this economic climate that junior engineers may not feel safe no matter how psychologically safe

your organization is and most are lacking in that to then ask those questions and to say what's they're stumbling on. It's very useful. How is learning changing now? Like is the platform some way to help junior developers or anyone learn AI more? I I went to a really interesting talk last week at Monkey GR in London uh which is a fantastic conference but the the one of

the talks from Sue Smith was about learning and she talked a little bit about muscle memory and the risk that we've kind of discussed already around if you only have senior engineers how do you end up with people who build in muscle memory because you have to start from junior you don't you don't emerge in the world as a senior engineer so she was talking about that

risk of not having junior junior junior engineers being trained. But then she flipped it around and talked about how actually with AI and LLMs learning is easier, more accessible. There are more ways that we can actually enable developers to learn. One of the things I was thinking about was where we can now with a platform easily spin up ephemeral environments. People can like easily spin something up,

try out something, tear it down again with low cost. That gives more of a safe space for developers to try out things, right? So you can give your junior developers a sandbox, an easy that isn't going to break the bank. It's not going to be expensive for the organization, but they can test something out and spin it down again without bringing down production or doing anything too

crazy. >> They're also see what you're writing. >> Oh, making notes. I was actually just writing about skills. So if you're approaching a new area or you're junior to a new area um you know certainly one of the things that we're looking at is how do we provide um standard skills or that that will embody the context and part you know the the learnings that perhaps as

a more um experienced engineer in a particular context that you've built up and if we can kind of encapsulate that in something like a skill and make that available does that help um junior engineers or engineers that are new to that environment to level up really quickly? ally and have maybe almost a little bit of a humanlike um pairing experience with um you know some sort of

prompt model whether it's claw or something else and that can actually be quite useful. I pair with mine quite a lot. >> What do you that's Yeah. What do you mean though by skills? because there's now we've got agentic skills and there's skills needed to get surfaced on a CV to get read by AI to get a job and then there's actual skills. So what are you

qualifying as skills for >> in that context? So I'm absolutely a junior in that area. So my terminology may not be right. So it's definitely kind of agentic skills. So, um, what we're kind of learning or trying to put together are textbased description files that are optimized for being read by whichever model you're potentially using. But the the sort of ways that we're using it is, you

know, if we've got one particular repo or a particular set of repos that always work together and there are interdependencies, it's modeling that, making that easy to, I suppose, reason about. >> That's a great use case. What other ways whether it's AI based or just platform based what are some of the different pieces or features that are common to a platform that can help junior >> Well

the main one would be surfacing all the information about um all the applications out there how they connect who owns them documentation for them previous incidents and run books. So it's about knowing where to go and getting the information so you don't have to go and ask someone who then points you to someone else and then someone else and then you know you end up back where

you started half the time. So one of the big um areas for me is being able to surface that information so it's easier to find for anyone and again it's not just juniors people moving around as well >> and AI >> and AI of course >> AI likes documentation. >> And can help write it. >> Yeah. I think it's also one of the reasons why we've seen

such a a rise in popularity of kind of developer portals. So things like backstage and port and cortex to Molly's point it's where you can provide a really easy interface human readable interface where you can see exactly what's available and like you said who owns it you can get to the documentation easily. I think that's why we've seen this increase because we are seeing more and more

complexity >> and for anyone coming in junior mid-level senior when you change job you may be a senior engineer but you may be very junior in that organization how do you find your way around what's currently existing having some visibility into that I think that's where having that centralized kind of viewpoint that platform is incredibly valuable >> but also that great context that um the concept of

a paved path or a golden path is one actually the platform has determined is a relatively frictionfree um approved way of doing something. And if you're coming to an area relatively new and there's a paved path, that's such an hopefully a relatively easier way of being successful quickly and without too much um friction. And and it's also it's a really good example of what someone or a

team has determined is the right way to do something. Um so there is a value judgment I think on a pave path or in a golden path. Um really good platforms also have an escape patch sometimes around those g paved paths. So if you do need to you know you do want to see what's happening under the hood or you want to kind of maybe experiment there

are those supported ways of doing it then that's great. That's great because yeah there was something that came up in the research of this is can a platform stilt just like AI the learning pathway of a junior developer because they just follow those paths which you want them to follow those paths but how do they continue to learn if things don't break or they don't have security

things which something I guess is very hard to convince your senior leadership team that we want them breaking things. >> Yeah, it's one of the things that does worry me a bit sometimes is you know I learned by things breaking right and troubleshooting them and getting into the logs and you know trying to fix the pipelines. So if they don't break then you know you are just

someone who writes the functionality and then if it goes wrong in production of course maybe it won't go in wrong in production because you've got all the right checks around it but it's highly unlikely right you're going to have problems so you know it's whether we can give them a safe environment like with yourself and the FM f FM environments you know so that they can try

and break stuff as well you kind of need to give them the safe space to be able to poke around a it and understand it because otherwise, you know, it's just always too good to be >> But it's also you got to think about what do you want to break? You know, do you want your CI/CD pipeline to break? You know, probably not. That's not necessarily where

you're you're going to learn. Um, you do potentially want your business logic to break in development or in the lower environments. And if platform can help you get there sooner without having to hit those blockers of you know how do you deploy something how do you spin up your you know your repository and getting all that boilerplate really quickly you are in a position to then experiment

much more freely. It's quite interesting. I was thinking about um whether now is the time that we have to do things like more intentional kind of education for our developers because we know that what the risk is if we don't have enough. Your your point about um intentionally breaking things like will we see the rise in popularity of kind of chaos engineering again because we haven't heard

about that for a little while which was popular once upon a time. One of the things Sue mentioned in her talk last week was about there's actually a lot of kind of information out there about how to do this intentional learning. Like we can't as an industry just assume that everything will be okay. There has to be some intentionality about how are we going to skill up

our developers and it's something we've got got to have a think about. But there's information out there. This is not a new thing to solve for, right? We just have to be intentional now that we're seeing more and more AI in the world. But all while teams are getting cut and platform teams are among the getting cut because last hired first fired situation or they like a

lot of technology teams sometimes struggle to prove their measurable value to business which that is something. So what we're talking about are things we all agree and we want to teach junior engineers and all. But if we're talking to the CEO or the board or the stockholders, how do we convince them that this investment is worthwhile when they're like, "Well, we can just use AI." >> Molly

made a really good point about the junior developers saving time for the senior developers, right? Um, and that's one way to position it because if you've got these very experienced, very knowledgeable people who are like full of the ideas and understand how things should be architected or how things should be designed, the junior developers will come in and take some of those lower level tasks. AI does

it too, but AI can't do everything. AI is, you know, generating code, but that's not all the programming is. So the pitching to kind of your CFO is we need these juniors because they are going to be enablers of the seniors. It's not just AI that's going to enable them. Something that came up this week because shock open source projects are mostly underfunded and yet your company's

success is built on top of them and ready to fall. uh is that if you use the platform to get a esbomb software bill of materials and you start recognizing what open source projects you're dependent on then you start looking how many maintainers some of those big maybe they're not the big graduated projects what if you get an apprenticeship an internship or a junior engineer on that

project think about that because it's certainly cheaper than if you have to deprecate a project that got abandoned and everyone's migrating from that project right now and you also then get to influence that project and maybe create influence that roadmap that's significantly cheaper way to get your security your your security patches gosh what is that over a year the average security patch for any open source project

so if you have someone or someone ones who are junior who become these stewards of this and they get work on these projects that could be owned by one of the big three or some of these big tech companies too and get a paycheck and maybe they're in a developing country because it's open source they can do it online. These are all really wonderful things that come

back to security which is really I think one of the big sellers to the senior leadership now. So what else are we seeing about platform plus AI plus junior developers? Does anyone have any experience where they're like well maybe a platform impede it or in this situation can platform help AI? Well, yeah, I think AI is like we said earlier is basically another junior engineer and to

your point they the coding is a small part of it knowing the context about your regulatory requirements or your environments you know how you set everything up that's really important and you don't get that from just letting AI go on your codebase that's where you need your pathways and your guardrails so yeah know I think platforms are really actually more critical now with AI are coming in

and hopefully well yeah they're going to save the day. >> That's what the Doora report found that was a strong finding that platform engineer is required because of AI. But let's not forget the juniors. >> But I suppose optimistically if we're looking at this um optimistically we're moving into in say into uncharted waters. They're perhaps not quite so uncharted, but we are the the junior engineers that

that are coming through are much more AI native than than perhaps the senior engineers that that are currently >> is that a problem? >> It isn't. I think it's a great opportunity and perhaps there are ways of of working particularly around the platform that we haven't even envvised yet and it takes um engineers with you know inquiring minds are curious but also with increased comfort with the

developments within the AI field to to build on that and to change and impact the way that engineering will look And then yeah just onboarding I guess we talk about this a lot but you need to onboard your AI and I don't know that a lot of organizations are yet it's a bit concerning actually because platform has accelerated the onboarding like I think Spotify is the one

that's touted that it to fully onboarded which is 10 PRs it was they went from 110 days I think to 10 or something because of a platform like that enables onboarding safely and inquisitively. So I guess maybe we should be onboarding AI or is this now just creating more overhead and overload for the platform team that's already overworked. >> I think it would be I mean Abby

who's from our team is doing a keynote tomorrow on on this topic actually but there's something that we are seeing which is more and more code being generated. people can go faster with AI, junior developers and senior and then the bottleneck is shifting into the platform. So we've been talking a lot about platform saving the day, but actually in some organizations, we're also hearing that the platform

and the underlying infrastructure itself is not quite ready for this massive influx of code that's coming and being able to ensure the quality and that human review if that's needed. And as I said at the start, platform should be an enabler. It should allow organizations to go faster. What it shouldn't be is a bottleneck. But we also wanted to have the guardrails and the safety and the

paved paths. So there's something about will we see organizations being developers managing agents and then platform teams managing agents in the platform. So that the speed at which the developers can produce code and features is then matched by the speed at which the platform teams can kind of release those features and manage those features day two and beyond. So I think to to Lena's point, we don't

know where things are going to end up. None of us have crystal ball, but I can certainly imagine a world where we're seeing like increase in one part of the stack and we now need to think about how can we use AI, how can we leverage AI more in the platform to kind of keep up with what's coming from our developers. >> So what we're saying is

we need more human beings in tech, not less because of AI specifically. think we can all agree with that. I hope you enjoyed this and learned and got something to take back to your senior leadership to advocate for the junior roles because if you are a senior, you should be not allowing your company to go quietly saying we don't want to hire any more junior roles and

you need to be the ones to advocate for it. So, thank you so much Paul and Molly and Lena and have a great day. Enjoy the rest of your CubeCon. >> Thank you.