Code with Jason
Code with Jason
299 - Eleni Konior, Senior Staff Software Engineer at Cisco Meraki
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
In this episode I talk with Eleni Konior about her path from economics to graphic design to programming, and how creative skills benefit technical work. We discuss building customer-focused features, the importance of assuming the customer's role, and AI in products beyond chatbots—like proactively surfacing recommendations based on user behavior.
Links:
A Different Kind Of Newsletter
SPEAKER_00Hey, it's Jason, host of the Code with Jason podcast. You're a developer. You like to listen to podcasts. You're listening to one right now. Maybe you like to read blogs and subscribe to email newsletters and stuff like that. Keep in touch. Email newsletters are a really nice way to keep on top of what's going on in the programming world. Except they're actually not. I don't know about you, but the last thing that I want to do after a long day of staring at the screen is sit there and stare at the screen some more. That's why I started a different kind of newsletter. It's a snail mail programming newsletter. That's right, I send an actual envelope in the mail containing a paper newsletter that you can hold in your hands. You can read it on your living room couch, at your kitchen table, in your bed, or in someone else's bed. And when they say, What are you doing in my bed? You can say, I'm reading Jason's newsletter. What does it look like? You might wonder what you might find in this snail mail programming newsletter. You can read about all kinds of programming topics like object-oriented programming, testing, DevOps, AI. Most of it's pretty technology agnostic. You can also read about other non-programming topics like philosophy, evolutionary theory, business, marketing, economics, psychology, music, cooking, history, geology, language, culture, robotics, and farming. The name of the newsletter is Nonsense Monthly. Here's what some of my readers are saying about it. Helmut Kobler from Los Angeles says thanks much for sending the newsletter. I got it about a week ago and read it on my sofa. It was a totally different experience than reading it on my computer or iPad. It felt more relaxed, more meaningful, something special and out of the ordinary. I'm sure that's what you were going for, so just wanted to let you know that you succeeded. Looking forward to more. Feels like a lost luxury. Chris Sonnier from Dickinson, Texas says, just finished reading my first nonsense monthly snail mail newsletter and truly enjoyed it. Something about holding a physical piece of paper that just feels good. Thank you for this. Can't wait for the next one. Dear listener, if you would like to get letters in the mail from yours truly every month, you can go sign up at nonsensemonthly.com. That's nonsensemonthly.com. I'll say it one more time. Nonsensemonthly.com. And now without further ado, here is today's episode. Hey, today I'm here with Lenny Konyor. Lenny, welcome.
SPEAKER_02Hey Jason, nice to see you today. And thank you so much for having me on your podcast.
SPEAKER_00Yeah, so you and I work together at Cisco Meraki. Um, you're not the first Merakian I've had on the podcast. I had Alan and Fido on the podcast before. Um and I've actually had a couple other of our colleagues on the show, but I think like before they actually came and worked here. Um anyway, can you tell me a little bit about like where you work in the company and stuff like that?
Role At Meraki And Team Scope
SPEAKER_02Yeah. Um, so as you said, we both work for Meraki. So that's one of the business units of Cisco, focused on the networking sector. And uh I work my team is called Foundational Features, so that's pretty vague, but um I'm a staff, a senior staff software engineer who gets to work with these features that just any other engineer at the company can also use. So like I can work with the API team, or I can build customer-facing features with uh the plain features team, or I shouldn't say plain, it's a great team, but um there's just so much breadth to it. Um it's one of the oldest teams at the company too, so or at least for Meraki, so it's pretty interesting to have all that legacy while also building Greenfield at times. And yeah, that's kind of where I sit um customer-facing features on dashboard. So if you're a customer of Meraki dashboards, you might have been using things that I've built before.
SPEAKER_00Yeah, and I've been there only about six months, but you've been there a while longer than I have. How long have you been at Meraki?
SPEAKER_02Yeah, I've been here three and a half years. I started April of 2022.
SPEAKER_00Yeah.
SPEAKER_02It's been it doesn't feel like I've been here that long now.
SPEAKER_00Um, and I looked at your LinkedIn profile, um, and you have like a really nice summary of your career and stuff like that. And I found it really interesting. Um, partly because it's just like interesting on its own, and partly because like there's some things that you and I have in common where it was like, oh, you play a stringed instrument and stuff like that. But um tell me a little bit about that. Like you went you went to college for I think economics. Um can maybe maybe we can start there if that sounds good.
Career Origins And Economics detour
SPEAKER_02Yeah, that sounds great. Um, yeah, so I I went to NYU and while I was there, like it a couple years had passed from the great financial crisis. And uh you get a couple years before you announce what your major is gonna be. And I had an environmental science professor who said recommended to everyone in the class to read The Economist because economists know everything. And I just found that so fascinating. And having like family in Greece who were significantly impacted by this financial event and seeing other people and my family impacted in America, um, I really just wanted to understand more about how the economy can have this great of an effect uh on people. And so I dived more into that and really enjoyed uh learning about economics. But yeah, it I'm not in economics. I'm not really using my degree. Um, I think by the time I was ready to graduate, like the idea had been to go to law school. And that didn't really feel like me. It just felt more of my beautiful immigrant parents who said be a doctor or a lawyer, and I definitely didn't want to be a doctor, so I picked the other option. Um, but you know, once I realized that was not something I wanted to do, I looked around and all my friends were kind of going into finance, and I definitely didn't want to do that either. I didn't want to just be like a suit at a desk. Um, not to say that that is something wrong. It just didn't seem like me. Um and so instead I actually pivoted and my my parents were entrepreneurs, so I launched my own graphic design company. I didn't really know much about graphic design, so I'm not sure why I was so bold as to create my own company, but I did it.
SPEAKER_00Wait, wait, if I if I can pause you there. Um just for context, you mentioned Greece and you mentioned your immigrant parents. Um you're you're from Greece. I just wanted to tie that together. Um, you were born in Greece and then your family moved to the fully American.
SPEAKER_02Oh, God. No, no, I'm first generation American. Yeah, yeah. My parents were born and raised and grew up all that in in Greece. My my mom uh emigrated when she was in her 20s. Um my dad had turned 18, and the next week later he was in Canada. Um he had an older brother who was there at the time, and then um he got an invitation to go to a wedding in New Jersey, and he went, he was like, Oh, I love this. So he stayed. And uh a few years later he met my mom because he was working with my uncle, and they kind of he kind of set them up and they met, they got engaged after three days, and then after three days. After three days, yeah. And then they got married the following April, so like four months later. Uh and yeah, they were together for over 30 years.
SPEAKER_00Wow. Okay, so I I I have to know more about this. Um, like what's what's the story behind this? They met and then three days later they got engaged. Like, that's obviously unusual. Um it is there more of a story there?
Family Roots And Lightning Engagement Story
SPEAKER_02Yeah, my my mom had come with my aunt um to so my mom is one of four. Um, my second oldest uncle from her side, so her brother, was working with my dad in New Jersey. And um, my aunt, who's younger than my mom, and her, they came for Christmas that year to come visit him. Uh, and so they had like tickets to go back to Greece after a few weeks. And um, I the story that I've heard is that my uncle had been telling my dad, you know, like, my sister's coming and I think she would be really great for you. And, you know, my dad was like, okay, like no promises, whatever. Um, and then the way that he had used to describe it was that he saw my mom. He was like, No, that's it. Like, I am proposing to this woman. Yeah. So he like talked to her a bunch of times and like they were going out of, you know, during those three days. And then he was like, Yep, I'm gonna propose to her. And so he like gets a bunch of flowers and he like hides them or whatever, and he sits with her alone and he says to her, um, okay, so what do you think? You want to get married?
SPEAKER_00Wow.
SPEAKER_01And she was like, Yeah.
SPEAKER_00Wow.
SPEAKER_02Yeah, really wild. Um, and so then he brought the flowers out and really excited, and they had uh they had a little celebration there, and then they had an engagement party shortly after, and then yeah, they had like I don't know, 200 or something people at their come to their wedding in like four months later. So yeah. And like I said, you know, it was three days, but it the marriage lasted over 30 years, so and it's still, you know, my my father passed away, so that's why they're not married. But I guess, you know. Okay, yeah.
SPEAKER_00Um that's that's really interesting. Um, like your dad was crazy enough to propose after three days, and your mom was crazy enough to say yes. Um there's some similar stuff in my family. Um, my parents, it's a different story why, but they both happened to be in Alaska at the same time. That's where they met. Um, and then I don't know how long it took, but like my mom got pregnant with me like fairly right away. Um, and then they got married later. And then me and my wife, we were together for like three months before we we got pregnant, and then we got married later. Yeah. So my parents had me when my mom was 19 and my dad was 20, and my wife's parents had her when they were like 16. Um and then my my wife's grandparents um let's see, my grandma had my wife's mom, sorry, my wife's grandma had my wife's mom when she was very young and not married also. So like on both sides, there's like these illegitimate babies and and the marriage happens after. Um so I have a theory that there's like a gene, you know, there's a gene for not being for being reckless with your reproductive material. Um, and and so that's how that happened.
SPEAKER_01Anyway, that's amazing.
SPEAKER_00Yeah.
SPEAKER_01Did you get to be at your parents' wedding?
SPEAKER_00Uh, good question. I don't know or remember. I assume so.
SPEAKER_01You have to find out.
SPEAKER_00Yeah.
SPEAKER_01I love that.
SPEAKER_00Um, okay. So you weren't born in Greece, you were born here. Um, and after college you started a graphic design company. What what impelled you to do this? Because economics and graphic design are obviously quite different. Um, what what was the what spurred that?
Pivot To Design And Self‑Taught Coding
SPEAKER_02Yeah, I one thing was that, and it's not the driving force, but I felt like I was doing so many technical things in college and not really getting to scratch the itch of my creativity. Um, so that was one thing. And then the second thing was while I was still in college, I had done uh some work with my dad. He was very involved in um an a Greek organization that um just gives back to people and you know, like scholarships and all of that stuff. So he was constantly having events or like needing magazines or something. And so I, you know, from no experience, I like developed a magazine for him. And um, from then it just like I kept growing those skills with more projects that I was doing. So my main focus was in that space of like giving back to the community. Um, I did some other like for-profit work that for like restaurants and stuff, I would do menus, but um yeah, it was just a way for me to have a creative outlet and try something to make my parents proud. Like, like I said, they were entrepreneurs, so I want to kind of kind of follow in their footsteps. Um, but yeah, I hit like around the two-year mark. I started saying to myself, okay, I feel more confident in my skill set. I really want to expand this business now. And every single idea I had had to do with code. And now, like I was a one-woman shop. Uh, I wasn't making that much money. And so I couldn't actually pay anybody to write any code for me and like get the business on fire to like just make more. Um, and so I decided like, okay, I could either not do these things, which doesn't sound like a good idea, or I can teach myself. So I found these free resources online. They I think they actually still exist, like the vote Viking Code Academy, or like they had something, a program called Odin or something, and it was so great. Uh, they had a Ruby option and a JavaScript option. I chose the JavaScript option. And uh it just went really in depth, talked about quality, all these things. And after that, I just really loved it. I started imagining myself like, how would you program something to like open a door automatically? You know, not that I actually wanted to do that personally, but I just kept thinking about solving everyday problems with code. And then I started dreaming about it. I was like, okay, I feel like this is something more than just like doing this for a business. So I went full into it. Um, I applied to the Grace Hopper program at Full Stack Academy boot camp in New York City. It's grown significantly since then. So I think they're remote and all of that. But at the time it was just in New York City, and that's an all-women's program, uh, which is the last time in my career that that will ever happen in.
SPEAKER_00Why is that?
SPEAKER_02Um because it was all women and like that was it. So and there only women, the only people in the program were women. Um, and you know, tech is obviously male dominated, so I'm usually the only woman in the room. Whereas when I was starting my career, I was one of many.
SPEAKER_00Oh, you mean that'll never happen again, not because you had a bad experience and you don't want it to happen again, but just because uh statistically it's unlikely that that that'll happen.
SPEAKER_02Yeah, correct. Uh no, it was a great experience. I I I don't know what the boot camp landscape is like now, so I don't know if I would recommend that path uh going forward. And AI has shifted so many things anyway. Like, I don't know, maybe go for to a bootcamp for AI uh if you're looking to get into software. But uh at the time it was, I would say, a perfect moment in the industry to to do that. And I had a great time just learning and diving into code and solving technical problems. It was really fun. And like, you know, back to that creativity. I felt like I got to do both at the same time. It wasn't like one or the other.
SPEAKER_00Yeah. Um yeah, and that's that's something I wanted to ask you about. Um, because a lot of people are like one or the other, they're either like predominantly right-brained or left brained. Um, but then some people are kind of blessed with both. Um, and it's it's not that common to encounter those sorts of people, but it seems like maybe you are one of those kind of people. Um, because you know, you play bass, so you're you're a musician, um, and you mentioned graphic design. So I'm kind of inferring that there's like some kind of an artistic side to you as well as the uh computer-y whatever side. Um am I kind of picking that up correctly?
Creativity Meets Engineering
SPEAKER_02Yeah, I feel like your words are very generous. Um, but yeah, I I don't really consider myself right or like predominantly right or left brained. Um, I think about everything that I do in from both lenses. And you know, there are times where there's stretches of time where things are very, very technical, and then there are stretches of time that where things feel very, very creative. But uh when I can get that man magic moment of both, I feel like that's my happiest point, really.
SPEAKER_00Yeah, same here. Um, oh, and I I also want to mention that I can tell you have um linguistic talent as well, just from reading the things that you've written and stuff like that, which is I think huge because there's I I know people who are like brilliant at math, but they like can't string together a coherent written sentence to save their life. And that's not I'm not being critical, that's just like you know, we come out how we come out, and you either have a brain like that or you don't. Um, and and so to have all those is like pretty rare, and and it allows you to do things that are kind of special, I think. Um like I've always been fascinated with people like Steve Jobs and Benjamin Franklin and Leonardo da Vinci because um they are not only I don't know, Ben Franklin was not only a scientist, but he was also uh a writer and a diplomat and a statesman, inventor, like all these different things, like these things that you tend not to find in the same package. Um and then with Steve Jobs and Apple, there's like this marriage of technology and art um which which Can produce things that are really incredible. I'm curious, do you think about do you think about those things very much? Like taking these two things that are usually not mixed and mixing the two?
SPEAKER_02Oh, that's a great question. I don't know that I always actively think about it. It's usually in conversations like this. Like I've had conversations with other colleagues of ours where we feel a lot of the times that computer science is more of an art than a science. You know, the beauty of writing elegant code that anybody can read and, you know, writing code that is for humans and not for computers, so that it's easier to maintain. Um, so I think about it maybe in the back of my head, but not like so consciously. But it's so interesting to me that you do pick these individuals to talk about because it really like it felt, especially in Benjamin Franklin's case, like it felt during that time when you read about history that the arts and sciences were so important, and that is how you had a well-rounded education. And we've definitely shifted in a different direction since then. But I I do feel like it's important to have a balance. Maybe you're not the person who like loves those things, and that's okay because it's good to have different interests, but uh, and not you personally, I'm just talking about the general you, but um, you know, it's important to still learn about them and at least cater to that side of your brain, even if you don't use it.
Design Principles Developers Should Know
SPEAKER_00Ray. Yeah, I think it's I personally really value having a well-rounded education and knowing a lot about a lot of different things, and not just having like islands of knowledge and understanding, but having like a tree of knowledge, if that makes sense. Like I I try to learn about everything and and see how everything can connect with everything else. Because then I think it's like your understanding is so much deeper if everything is connected than if you have these just disconnected islands of knowledge.
SPEAKER_02Yeah, totally agree. And usually you can find a thread between things. And you know, I I had a professor, uh, a different professor, he was my con law professor, Professor Raj thing, uh, if he's listening ever. Hello. But yeah, he he said, write about the things that you love. And so he would give us tasks about, you know, some kind of constitutional uh case that we needed to read about and and write about. And he was like, well, you know, I gave you this assignment, but find a way to write about something that you love that has to do with this case. And so, like learning about things and that you're interested in, and then finding ways to connect, it just enriches everything that you're doing and learning about, and and you can like teach it to others, which is so exciting. And yeah, I just find that so fascinating. I love thinking about it, like how you said as a tree, definitely with different branches and and leaves.
SPEAKER_00Yeah, yeah, I found that very useful. Um okay, so econ in college, then graphic design. Um, and I'm curious with the gr with with your experiences in graphic design, and then you went to programming. Um, do you feel like your your interest in and experiences with graphic design have benefited you in programming? Like have you used that stuff in programming?
SPEAKER_02I actually have, which I didn't ever expect. Uh, I am a full stack engineer, so there are again stretches of time where I'm doing like very heavy back end work and not looking at the front end, and then other stretches of time where I'm doing like very front-end focused work. And I usually, if I've done one for too long, I need to flip back to the other to keep myself sane. But whenever I'm doing front-end work, I I can't help myself. I'm always talking to the UX designers and like, hey, like, have we thought about this? Like, what about this for the customer experience? And, you know, I think this might be better suited elsewhere. And, you know, I always preface it as like, I am not a UX designer. Like, you can totally tell me to like fuck right off. But I like I just can't help myself that I always think about the customer and always think about like the layout of things and making things simple because of that graphic design background. So luckily people have indulged me and love, at least they tell me they love it.
SPEAKER_01I don't know if they really do, but they told me that they love it when I give them the feedback and like basically pair with them on the UX design.
SPEAKER_02So yeah, yeah, it's exciting.
Building For Users, Not Just Requests
SPEAKER_00Um, yeah, I think that's really powerful. Like when a programmer can become competent in design, then that it's it's just when all those skills and abilities can be contained within one person, then then it's it's so powerful. Sorry. Oh, it's okay. I think we're back. Um but yeah, uh I I found this really great book that I've talked about on the podcast a lot called The Non-Designer's Design Book. Um and it described a few principles that I've that have stuck with me ever since then. I'll just really briefly mention them. Um like contrast, repetition, um, alignment, and proximity. And just those few principles have like gone such a long way. And then I've read a couple like um usability testing books, books like um uh Don't Make Me Think by Steve Krug. That one was really and and just these few things, like learning just a little bit about UX has gone such a long way, and I wish that every programmer would study these things, because then you know software might be software is like notoriously terrible, and so and a lot of it is because developers just aren't trained in these in these design principles, um, but a little bit can go a long way.
SPEAKER_02Yeah, I totally agree. You know, I actually saw a LinkedIn post the other day where someone was a new hire at a different company, and uh like he has a wealth of experience, but he's new to the company, he's only been there for four weeks, so he's not expected to be on call. But this like something happened and everything was on fire, and uh this person who was paging people didn't find anybody else to page, so paged him. And he was like, I have never used this platform before I just started. Um he's like, you know what? I'm gonna just try it and figure it out. And uh his post at the end was, you know, you know, I was able to figure it out because this tool was built so well that it made sense. And I think it was an internal tool. Um, but you know, I just that really sticks with me reading that because there's no better thing to hear as an engineer, like when someone is using your product that it actually did what you wanted it to do and helped them get something done.
SPEAKER_03Yeah.
SPEAKER_02And he didn't need to read a manual to figure it out, it just was sensical. So that's just so elegant in terms of programming to me.
SPEAKER_00Yeah, that's really nice. Um, so how do you like I'm curious how you think about this? How do you ensure or or try to ensure that uh an interface will be intuitive and easy to use?
Usability Testing And Iteration
SPEAKER_02Oh, that's a great question. Um I don't know that I have all of the answers. You know, I try to assume the role of the customer. I think that's very important. If you're building tools and only thinking about how you're gonna maintain them, you're never gonna build a product that is actually helpful to people because you can make things maintainable for yourself, but also like wicked fast and like really awesome for the customer. So I try to always assume that role first. And then when it comes to UX design, I'll try to, you know, obviously it isn't built yet when the by the time I'm looking at something and like it's just a Figma screen. But um, you know, if you're lucky, there's some kind of interaction built in from the UX designer. If not, you just kind of try and follow it. And so I try to envision, you know, if I'm doing an action that the team is supposed to build on the page, me as a customer, is it actually going to help me do this, do it quickly and without error? Uh, because you know, that's another thing is if you build something and then the customer is allowed to do it and it messes up everything for them, especially in networking, like the industry that you and I are in for tech, it's it can go terribly wrong. Um, so you want to make sure that you have the guardrails in place to help the customer while also giving them permission to supersede those guardrails if if need be. And so I I think it always comes back to just trying to be like the customer. Um, I've you know, when I was working at Bloomberg, that was my first like full-time software role out of the boot camp. Um I learned that lesson very quickly. And you know, that was fintech. And so if you made a mistake in the code, that could cost like millions of dollars, and that is always very scary. Um, so you want to kind of assume the role of a trader almost in a lot of the things that at least I was building. You know, traders aren't the only people who use Bloomberg, but um, it's really important to deeply understand your customer base and even pretending to be them, insofar as like creating your own setup and trying to execute on these things, uh, whether it be a trade or like setting up a network in Meraki. Um, I I find that is very helpful to getting your frame of mind in the right place for building features for them.
SPEAKER_00Yeah, it makes me really happy to hear you say that. Um, it's it's sadly not a common practice to like try to understand the people who are using your product as crazy as that sounds. Um but a lot of people will like they see it as their job to to build the things that their users are asking for. But I don't think that's really quite it. Like uh I think your job is to understand your customers' world and their job and everything, and and you figure out what they need in order to do their job, um, not just listen to what they ask for, um, because what they ask for may or may not be what they actually need. Better to deeply understand, like you said, what they need to do and then build based off that.
SPEAKER_02Oh, yeah, 100%. You know, I try to subscribe to mentality that the customer is always right, but sometimes they don't know what they need. And so it's important to, you know, you can give them what they want always, but if you can take it a step further and be like, well, this is your workflow, and I can pick up on the patterns of your workflow, and like this would actually make your life so much better. Um, I think building those kinds of features is where this like the magic really happens. And you know, it's always great when you get the positive feedback and reinforcement from the customer, like, I had no idea that I even needed this in my life, and you built it for me, and that's awesome. Sometimes you don't always get there, and that's okay. You learn from it for the next time, but yeah, it's definitely a balance between need and one.
SPEAKER_00Um and and I think testing is a big part of it too, uh usability testing. Um, I I used to do a lot of this when I was working at an in-office job, and now I don't even build like user-facing, I don't even really do any programming anymore um at at Meraki anyway. Um but um it's like the only way to know if your interface is gonna do the trick is to test it and put it in front of somebody. And the the key thing that I learned is like don't put it in front of somebody and ask, like, what do you think of this? Um because people will always say, like, oh, looks great, you know, they'll be polite. Um, but then the second you like leave and give it to them, everything falls apart. So you have to give it to them and say, Here, use this in order to do such and such task, because they can't fake that. Um, and and what I found is that like basically 100% of the time, my first attempt at making something highly usable is way off. And like as soon as the rubber meets the road, like it's crash and burn. And I'm like, oh wow, okay, I didn't realize that you would like think that that means that, and I have to like fundamentally rework it. Uh, you really can't predict how people are gonna use the thing that you build. Um, and I don't know if if if we do that at Meraki very much, the usability testing thing. It's it's very rare, especially now that everybody's remote, but I I I think it's the the main thing that like ensures a usable interface.
Beyond Hype: Practical AI In Products
SPEAKER_02Yeah, we have uh a UX research uh sector of the company that does these things, and like they'll also ask really thoughtful questions of the customer, like you know, that need versus want. Um and but what we're lacking, and I'm sure like most companies do this, and you know, this is just my opinion, I have to say that. Um it would be better if we were able to iterate more quickly. And a lot of teams have started to do that more, which is really wonderful to see, but it's not streamlined across every team. And so if you can show the customer something when you're early on in the development process, it's a workable product, but it's you know, maybe a step below MVP. Like you tell the customer up front, this isn't perfect, but we want to make sure that this is what you need. And they, you know, do that usability testing, like you said, like please use it to do XYZ. I feel like that is the best kind of feedback that you can get because you can spend multiple quarters on something and then deliver a product and like cool, high fives, everyone. We did the thing, and then the customer doesn't use it, and that is luckily that's not happened to me. I like that would just give me so such panic if like no one was using the things that we were building, and you know, it I've had customers say, like, this doesn't work for me, and that's great, like we can iterate. Um, but to by the time you're done with it, if you haven't gotten that that feedback from the customer who you are building things for, then it's just a long and arduous road. You're likely never to come back to that feature, and then the customer's upset.
SPEAKER_00And yeah, yeah, and what a waste. Like, I really don't buy that like lean startup thing of like their idea of operating scientifically is to like make hypotheses about what you should build and then build that for people, and then oh, maybe it's not what they needed. Um, and so you build something else. It's like, shouldn't you try to figure out that before? Like do the research before and figure out what's needed so that by the time you're investing the the time into building it, um, you're not making a wild speculation. Maybe maybe it's still kind of a a bet, you know. You you can never be absolutely certain, but there's a difference between buying a lotto ticket and uh I don't know what the other analogy is, but like you can make a much safer bet if you do some research first. Um so that's that's good to hear that there's that there's research being done. Um let's see. We were gonna talk about AI, but we haven't gotten into that at all. Um we have we have a little bit of time left, so maybe we can get at least a little bit into the AI stuff. So obviously, you know, as you and I talked about pre-show, uh everybody's talking about AI-related stuff, but a lot of it is like on the developer side. And you mentioned some interesting things about like on the feature side of um incorporating AI into products. Can you tell us a bit about what you're thinking about?
Recommendations, Classic AI, And Hack Week
SPEAKER_02Yeah, absolutely. Um, so uh this started for me probably around the same time that ChatGPT launched. So ChatGPT came out almost, what is it, two years ago or something? It came out in 2023, end of November. Um, and a few months later, I my product manager asked me to join a virtual squad uh that he was setting up, having these like awesome projects that he wanted to build before Cisco Live US that year. Um, and so you know, I naturally was like, cool, let's do this, let's build it together. I got to lead this project. And it was the first project that I got to build that had, you know, some intelligence and AI in it. And uh, you know, typically in the software process, like whatever the design process, whatever I knew before Chat GPT, was, you know, you hopefully are doing that research for customers, you're talking to them so that you know exactly what they want or need. Um, you know, obviously that would shift over time because you know, you're getting snapshots in time, and maybe that has shifted by the time the product launches, but uh you have a decent idea of what they are kind of looking for. And it can almost be like rinse and repeat, you know. So design the thing, build the thing, ship the thing, get the feedback from the customer and keep going. But how do you take it, you know, almost three steps further now that we have these tools to just float up to the surface? Like I, if you understand the customer really well, you know their workflows, how can you not only surface any potential problems, especially in networking, right? Like, how can I tell the customer what problems do you have with your network or networks and not only what problems, but how can you go and fix them? And ideally, there's like the third part of if you just click here, we'll fix it for you, right? Because we have it's the code base, right? We have all of the tools that we need. And, you know, effectively that might sound like automation on steroids, and that's kind of where we're at in, I would say, in the AI landscape right now, but having that as a customer-facing feature and not necessarily charging them for that utility, that just makes them so much happier because now they're getting so much more done faster and that gains trust, and then you can do more things. And so it's really interesting to think about like any kind of AI feature that a customer has. And it's not just you know the developer tools that we don't talk about, but how can I just make that workflow like really seamless, really snappy, and you know, very simple too?
SPEAKER_00Yeah, yeah, it's a lot of um AI features in products that I've seen, and I'm curious if you feel the same way. It feels very like um tacked on, and like it uh the the idea seems to be like how can we uh uh shoehorn A high into our product somehow and do it like in the next few weeks? Like, give me something, developers, and it feels like that's how it how it came out. Um, but it feels like there's a lot more opportunity than these like superficial things that have been done so far.
SPEAKER_02Yeah, and there's so many different levels too. Like we can build things that are like connecting to LLMs and give you know customers a a summary of what is going on or um you know. Some recommendations that we might give them based, and that's kind of like the prompt engineering approach. So they don't interact directly with the LLM, but by using the product, it's like under the hood. So you can't chat with it. You just kind of get the results. But then you can also do things that are so much simpler that might not look like AI, but they're under the hood. They're like classic AI algorithms. My team actually just did that. We had a hack week. So I I prepared three projects for people and like did like mini PRDs for them, um, for the people on the team. And I said, okay, pick your favorites. We're gonna divide into teams and do this for a week. And we got really far in the one week, so that was really awesome. Um, but one of the projects was, you know, we're building, I don't, I can't say because it's not released yet, but um, it's effectively like a list and you can do a bunch of actions on it. And uh, you know, we don't really expect our customers to actually sift through the lists and look through everything. So it would be better if they can surface the information that they need much quicker. But if they did want to, you know, page through all of the data that they they should be able to. Um, but you know, one of the actions is to put one thing into some other thing. And again, apologies for being vague. I can't really talk about it. But um wouldn't it be better if when they're doing that, we can actually tell them that based on their other configurations, we would recommend that they put that thing into a specific place. And they can override that, right? Like we're not telling them that it's mandatory. But just that one simple thing is like light years ahead of just saying, okay, well, I provided you all the data. Now you have to sift through it yourself. No, we don't need to do that. We can say, like, hey, if you want to sift through it, you can, but you're gonna click this, and then once you launch this thing to click it, you're gonna see our recommendation in there. And I just find that really special to be able to do that. It can be so simple, but it it really does make a difference and an impact on the customer experience.
Data Shaping, Prompting, And Reliability
SPEAKER_00Yeah. Yeah, I think that's a really good application of it. Like, there's a general pattern that I've been thinking about a lot and using sometimes where it's basically like it, it's it's nothing all that spectacularly novel at all. It's just like you have your SQL database, you have an LLM, like, why not feed the schema to the LLM and then say, okay, we're we're gonna give you some natural language questions and just say things like uh who are my who are my top three most active customers over the last two weeks? And then the LLM can take that question and take the schema and make an SQL query out of that and then give you the result. Um I don't know why like we're not seeing that more. Um, because it does it seems like one of those ideas that is just like it's like a good idea waiting to happen, you know? And I'm sure a lot of people have independently have that have that same ide same idea. Um, I don't see it being implemented a lot, but that seems similar to if I understand what you're describing. It sounds kind of similar. Like you have the data, you have the question, and it's kind of a fuzzy thing where you you can't just do like a text search or something like that. Um, but you're you're making it convenient for the user to get at the information that you need. Is that the idea?
SPEAKER_02Yeah, you know, like the other alternative would be if you're talking about the UX, you allow the customer to search for something so that you get the results back. And this just takes it a step further. It's like you don't even need to search. I'm telling you exactly based on the things that you've done already where you are likely to put this again.
SPEAKER_00Oh, interesting.
SPEAKER_02Um, and yeah, and you can go ahead and and still do the search if you want to and change like pick a different route if you wanted to. Like that's the beauty of the design for this particular project that we had done is um or sub-project of an overarching project. But you know, you just give you give the customer the options, but you also just like float that information immediately so that they don't have to spend more time doing that search themselves or clicking through things.
SPEAKER_00So yeah, that's a really interesting thing.
SPEAKER_02It seems like automation on steroids, but uh it's just a more elegant experience, I think, by doing that. And there's that's just only one example, right? Like the Cisco Live project that I mentioned before, that that was analyzing packet captures. And um I didn't even know what a packet capture was before 2024. Um, I knew it existed. I didn't know what it did. Um, and so being able to actually tell someone, especially someone who might not necessarily know how to read such a file because it's actually very technical. Um, and there's like a separate software that you need to download to be able to click through all the different pieces of a PCAP and see what is happening. Like, what if we also embed that in our own application and and further than that, tell them exactly what is wrong. And then again, a step further, like, okay, we've told you what's wrong. If you fix this like ABC, everything's gonna connect again beautifully, and you're back to normal and you're just gonna see green everywhere. Um, so that kind of experience I think is what we should be focusing on a lot more. And like, of course, every company wants to have like their chat bot. And I do think that there's value in that, but that is not the only thing that we should be building as engineers.
Bass, Puzzles, And Parallel Paths
SPEAKER_00Yeah. Okay, so tell me if I'm understanding this right, or I'll explain like how I think about it, and maybe tell me if if it matches what you're thinking. Um so again, a lot of what I've seen, what I've seen with AI is just like, oh, let's sprinkle some AI on it. Like take the existing thing and try to plug AI directly into the existing thing as it currently is. Um, but it seems to me like there's so much more opportunity if we take the existing thing and then build some layers on top of it, like specifically targeting the AI. Um for example, um, well, I was gonna give a Meraki example, but only you would understand that, and nobody else. Um but like with with with this querying your database thing, um, maybe for example, it wouldn't be so easy to it wouldn't be so easy for the AI to take your question and formulate a meaningful query like on the raw data. So maybe you can make like an SQL view or something like that, um, so that you kind of constrain it and narrow it a little bit and say, like, okay, here's the subset of data that you should be interested in, LLM. And so you can you can make your queries targeted at this. So again, not just taking the raw material, but building up so that the AI has something to connect to, so that overall that can provide provide a much more valuable feature. And it sounds like maybe that's that's a similar thing to what you're thinking, um, but I don't know if I'm right about that.
SPEAKER_02No, I totally agree. Um, you know, it doesn't necessarily have to be like in this example, we're talking about a SQL database, but it could be anything that is like parsing data. And how do you get the correct data or the most helpful data to an LLM in order for them for it to give you something back that's meaningful? Um, and that's actually what we should be doing, right? We get it's good to start off, you know, as the first pass, like just send everything that you might need. And of course, you know, you want to always protect the customer and their data, like strip it of PII and all of that. Like, I I'm a firm believer, and I also learned this at Bloomberg, like the data does not belong to us just because we're engineers, it belongs to the customer. And so you have to be very careful what you're doing and like what you're training it on and all that stuff. But if you give it the right data, and of course you have permission from the customer to do that and like all the things that come with it, then um, yeah, you get better results back. But you can iterate, like I said, and and try it out and see what works better. Um, I do call that prompt engineering because essentially you're giving the LLM a prompt and you're saying, okay, I want you to do these things, and this is the data that I'm gonna give it, give to you, and then see what results it spits out and try to tweak it from there. And that could be a better subset of data, that can be a better prompt. Uh, it could be multiple things. But yeah, it's it's about finding that that meaningful bit, which is kind of actually hard because AI is not really deterministic all the time. Like it's probabilistic. Um, and you know, back to that hack week example, I was working on a different project. Um, and I saw the LLM, like at one point it was infinite looping, and it I gave it the same data every single time. And it was giving me different results back every single time. And that's not great, right? Like you want to get it into a spot where most of the time you're gonna get the same results. Um, so finding out ways to do that, and I know you talk about testing a lot on your podcast, which I love. Um, so finding ways to actually test it so it is somewhat deterministic, I think is gonna propel those features even more forward than they are today.
Where To Find Lenny Online
SPEAKER_00Yeah, I totally agree. Um, there's there's one more question that I want to ask you before we have to wrap up. Um, and it is bass related. So I'm I'm just curious, what got you into playing the bass? How long have you been playing bass? All that stuff.
SPEAKER_02Yeah, um, I started playing the bass. I was in middle school. And I don't know what struck me, but I had always wanted to learn some kind of guitar. Um and I remember being in gym, and the music director of the school came to talk to us and was like, I have spots open and I need another bass bassist or whatever. And I was like, huh, okay, I don't know anything about that, but I'm at school, I'll learn, you know. So I I joined band and I got a bass. I still have my bass. I love it so much. Um, like that first band I ever got. I still have that one. Um, and yeah, I I was slapping the bass. I was learning all the notes and like how to read sheet music, and it was really exciting. Um, by the time I got into high school, I stopped band. I wanted to, I was so focused on like figuring out what I wanted to do with my life. So I tried to do different um elective rotations every like quarter that we had them. Um so I stopped doing because I knew I didn't want to be a musician professionally, which like I don't know, maybe I should have. But yeah, that's kind of how I got into it. And then uh years ago, I think it was during the pandemic, um, I found like basically guitar hero, but with real instruments. I forget what it's called, but uh you literally plug in the bass to like a PS controller, a PlayStation controller, and playing bass, and it's like it looks like almost like a guitar hero interface, but it's your it's a real like instrument. So yeah, uh I don't get to play it as much as I'd like, but that's kind of how I got into it. And I had a friend who was doing the bass with me in in middle school, so it didn't feel like I was just doing it by myself.
SPEAKER_00Yeah, interesting. Yeah, I I no longer own a bass. I I have, I'm primarily a guitar player, but I was in a band, a power trio, where I was the bass player slash singer, which is kind of an unusual combo. Um, but there's there's something that I just really like about the bass. There, it's like a a puzzle of like, okay, you have some chord progression and you have to come up with a bass line that follows the chord progression but is like interesting on its own. There's something almost mathematical about that puzzle to me that I really like. And it's just like a fun instrument with like those big thick strings. There's something about that that really appeals to me.
SPEAKER_02Me too. Wow, I'm so happy that you called it a puzzle. I never thought about it that way, but it totally makes sense. Uh, I I should have realized back then that I would become a software engineer because I love puzzles.
SPEAKER_01So bass guitar programming, you're like constantly solving puzzles, and there it is. I love that. That's cool. How long did you do the the band for?
SPEAKER_00Oh, this was probably only like a year or so. There were there was a band that which was my main band. Um we were called the moms, and that was that was like through mid-high school into college a little bit. And then this other band, it was like it, it was how do I describe it? I don't know. I don't want to be like mean to the other guys in the band, but it was like it came together out of like it was just like who I don't know. I don't know how to put it, but anyway, the moms was like that. And then this other band we were called, we were called the sexual neighbors. Um and and that lasted for like a year. We always played um in our underwear, um and so that was uh it's on brand. Yeah, yeah, and that lasted for about a year, and then I moved away. I was living in Kalamazoo, Michigan, um, going to college there, and then I moved to Austin, Texas. Um, and then pretty much no more music after that. I've tried to join bands and create bands after that. It has never really worked. After you have like a kid, uh after you have kids and a real job and stuff like that, it's the rock and roll lifestyle is kind of closed off to you after that.
SPEAKER_01Yeah, I bet.
SPEAKER_02Wow, what a what a great thing to be able to tell your kids though that you were in these bands, and like I just love that, you know, like nothing is linear in life, and it's so cool to that like the fact that we have so many overlaps too is really fascinating to me.
SPEAKER_00And I see your piano behind you, so oh yeah, I was gonna ask, like, do you do you ever dabble in any other instruments or just bass?
SPEAKER_02No, just bass. I do have two guitars, but I've barely played them. Um and I did the flute before I did the bass, and yeah, like I was decent at it, but I was never gonna have that as like my main thing. So I just found the bass to be so much more fun. Um didn't want to take myself too seriously when I was learning something really fun.
SPEAKER_00Um yeah, this piano behind me is pretty much just a decoration at this point. I never play it. Um yeah, uh so very last question for you. Um if if people want to go online and learn more about you and what you're up to, where can they go?
SPEAKER_02Oh yeah, uh I do have a personal website where I put a bunch of things on there, like things I write, which I haven't done in quite some time, so I probably should, and uh things that I use as a software engineer. Um, also like my resume is on there, and that is datgreekchick.com. D A T Greekchick.com.
SPEAKER_00All right. We'll put that in the show notes. And Eleni, thanks so much for coming on the show.
SPEAKER_02Thanks so much for having me, Jason. This is really awesome.