Remote Work, Tech Careers, and AI Insights with Derek Yang

In this episode of What’s the Startup?!, we sit down with Derek Yang, a Senior Backend Engineer who’s navigated the twists and turns of remote work, career pivots, and the evolving world of tech. From making the leap into coding through a bootcamp to finding his way into the Sprocket community, Derek’s story is all about embracing change, learning on the fly, and staying ahead of the curve in a fast-moving industry.

We dive into what it really takes to transition into tech, how to leverage coworking spaces like Sprocket for connection and support, and Derek’s perspective on the AI revolution. Whether you’re already in tech or just curious about taking the leap, Derek’s insights will inspire you to carve your own path. This episode is packed with real talk about navigating your career, making the most of remote work, and why building relationships is key to staying motivated and driven in any field.

Tune in for practical advice, a bit of skepticism around AI hype, and plenty of inspiration to help you move forward with confidence, no matter where you’re at in your journey.

Thank you for tuning in to this episode of the Sprocket Podcast! If you’re ready to dive into the world of startups and innovation, visit us online at Sprocket Paducah to learn more about our mission and how we support entrepreneurs like you.

Got a business idea? Don’t miss out on our West Kentucky Innovation Challenge, where you can test your concept, build a prototype, and compete for prizes that can launch your startup to new heights.

Stay connected and join our growing community on Instagram for the latest updates, inspiration, and behind-the-scenes looks at what’s happening at Sprocket.

Let’s turn your ideas into reality—together!

Sprocket is proud to be supported by Team Kentucky, the Commonwealth's Cabinet for Economic Development. Learn more about their initiatives and resources at ced.ky.gov.

Transcript

0:47

Welcome to What's the Startup, Derek. Thank you. Thank you for having me. Do you want to tell us a little bit about who you are, why you're here, and why Sprocket? Sprocket. Yeah, sure. So I'm going to try to shorten the story as much as possible. I, my name is Derek. I grew up in Southern California. Went to school in New York, lived there until 2020. The pandemic hits. Where'd you go in New York? I went to Columbia. Cool. No. So the pandemic hits it's 2020. I was working for a company called Birchbox at the time. And we went full remotes. My parents called me and said, Hey, why don't you come out here? They were already out here. They started a fishery out in Wycliffe in Ballard County. Oh, wow. They were like, Why don't you come down and spend some time with us? You're already working remote anyways, were you looking at Paducah in particular? It's where they were. Yeah, you're like, this is the spot. How did you find Sprocket out in the wild? Just looking like working remote gets old after a while when you're just working from home. So I was always looking for other places to go. And I'd hoped that there'd be a coworking space and there wasn't initially in 2020, it wasn't a thing yet, but lo and behold, and I think 2022 started, it became a thing and I came in here. Were you one of the first, like the inaugural kind of member? I was in here pretty soon after they opened the coworking space. I was here for opening night. Oh, cool. Yeah. Cool, cool. Yeah I know that you have been co working here, but as far as like your day to day work, what are you working on? Like, how did you shift over into this part of tech? Yeah, so a little bit of context. I am a Senior back end engineer. I primarily work in web development. I, my stack of choice is Python, Django web framework, and then, various tools for infrastructure deployments and etc. I didn't major in computer science in college. I was actually a physics and then a financial economics major. Nice. Just want financial economics just because I was tired of school. I wanted to graduate faster, and I didn't see you know a PhD in my future and like continuing like scientific work because that's I At the time I thought that's what you do with a physics degree. There's not much you can do with a bachelor's You know, and, yeah, I just didn't, I didn't feel like I was good enough at it nor did I have as much interest in it than I initially thought. So I switched majors, and then I graduated, and I think, like a lot of people, I took the first job that made me an offer, because I was getting kicked out of student housing soon, I needed to find something, and no landlord was going to rent to me if I didn't have a job. So I took a marketing job at a marketing agency. Great people. It's a marketing agency that specializes in industrial manufacturers. But I found that I didn't love the work, so after staying there for about a year, I saw an ad online for a coding boot camp, and this was 2016 time frame, and I decided to take the leap. I do think moving to a different area, a place, a space like Sprocket provides an invaluable service because it allows, it allowed me to meet people. I don't work directly with anyone here, really. My work is mostly centered around my company. I don't have any sort of professional links to this community. But it's been nice just to get to know people and just to have a sense of an office again. Yeah. Yeah. So it's like coworkers without the baggage. Exactly. Yeah. It was like you only get the good stuff. Yeah. Really cool. Yeah. That's a good way of looking at it. Yeah. So you get the water cooler talk and you get the camaraderie and the networking and everything, but maybe you don't have like that. It's cool because you can jump out of what you're doing heads down, like day to day and then discuss how other people are solving problems in this space. I've done a little, I've experienced that too around here. So that's a super cool thing. I know there are a few others in the space like that, but when you come in does it, what's different from working from home? Like when you come in to Sprocket in the offices and take advantage of some of the member services. Yeah. So I think, one of the major differences is just the frame shift. When I come into Sprocket, I'm dressed like I'm about to, do a day's work. My mind is more refreshed. Usually I try to hit the gym before coming in. It's just a much different sort of mental state than rolling out of bed and into my first meeting with the camera off. Yeah. I've done a lot of that too. Yeah. It was fun for, a month or two during the pandemic when we all had to stay home. But yeah, you start scaling the walls after a while. Yeah, exactly. And to be able to have a healthy mix of maybe a little bit of both really works nicely. I think specifically in tech, is there anything like. I'm sure it will only speed up from here. Like the advent of AI and like the integration of however we code and build in the future. Web architect, web development, whatever, whether AI agents or not. Yeah. What do you think that looks like for, You personally and more broadly across tech. Sure, yeah, so AI is the hot topic of the time, right? It's definitely, we're in the middle of this it's being hyped so much. And I don't completely disagree with all of the hype. I think it is a very amazing transformational technology. When I first played around with ChatGPT, I think, Early last year. I was like, oh man, I bet I had like electrician Apprenticeships bookmark like in case this goes south. I need to find a new line of work. Yeah But really so far. It's just been a very helpful aid It hasn't it's like having on hands a very motivated Intern to go look up questions for you and very motivated. Yeah, And they might not get everything right all the time. You might have to, you'll have to double check their work. But, tireless, really polite. Yeah. Which, I'm always yelling at my chat GPT No, you're wrong. Yeah. Tell me the right way. It will, what some people have called Hallucinate. Yeah. And invent things that aren't necessarily true. Yeah. Which is odd from like a logic machine or like from typically what you feed into a computer. Like it, it gives you a very let me start it again. Sorry. So what do I really want to ask you about AI? Think about this. So with AI in particular, I've seen a lot of like agent or people build agents to go and do different roles. To me at a glance, that looks like a very powerful like enterprise tool to be able to do at this like multifactorial, like many different agents working together maybe not always out in the open, but able to do it between each other. So how do you, so I get the gimmicky part and then like people really hyping up AI and then stock price go up and all that stuff, but separate from that, the AI in in terms of building things do you think we're ahead of, or we have in store some exponential growth there with agents? I don't, so I'm a little I think personally I'm more of an AI or I'm not an AI skeptic, so I want to define the conversation. When we say ai right now, we're talking about large language models, right? Yeah. Like the generative ai which is like what chat GPT is, that's really a subset of a, of machine learning, which is a subset of ai. now. I think a lot of people conflate chat GPT like LLMs to eventually reach what is known as AGI, Artificial General Intelligence. And that is an intelligence that can actually reason and solve problems. That's not what LLMs do though. LLMs are just averaging things. They're trying to predict, like in the case of a language model like GPT, it's trying to predict the next. Word in a sequence after having consumed a massive astronomical amount of training data. It fig it knows how to Compute the values, the statistical value of certain words in relation to each other. So it's able to calculate, okay, based on this input, statistically, what should be the next sequence of words. And when you, when we, what we find is that over a large enough training, training data sets, one of the emergent qualities of that is that what it's spitting back at us sounds really good. Really conversational and it's right like 80 percent of the time, so it's an amazing tool. But I don't think that is what's going to lead to AGI because at the end of it, the model is not aware technically of anything you're asking it. It's just aware of this sort of mathematical relationship between different terms. And that's, yeah, that, that's really it. It's like a weather model that's only looking at historical trends and trying to plot the next day. It's like weather without looking at anything else. It's just looking at the trends. So intelligence in, in that sense is a, not a misnomer, but it's a little bit gratuitous. It's a little bit like, it's a, it's not exactly the most advanced machine learning per se. And this is good because I'm pretty much a novice on. the mechanics of it. So I'm really glad to be talking to you about it. Yeah. When we're talking about the large language models it's a very predictive, whether it's generative, like a picture, like mid journey, you're predicting the next pixel or whatever the mechanics are. It, the emergent, I like how you said that emergent quality of it is that it actually works. And I, I think that's, I don't know if mysteries are the right word, but that's, what's coming out of. A lot of these enterprises who are building the LLMs out at scale is that it's working. Maybe it's not right to say that nobody knows exactly why, but it is a very functional tool at the moment. It's, there is not, it's not like a wear, it's not, it doesn't have a thinking brain, so to speak. At least it's, we can't perceive that. That's my understanding. Yeah, it's exactly where it keeps keeping us where it wants us. Yeah. Before Skynet turns on. But I think there's nothing that it is like to be a large language model. So there's no experience. It's not thinking. So it's predicting the next word. And that just so happens to be very helpful. And if you're doing this marketing analysis or if you're doing this Breakdown of hey, write this business plan for me. Yeah. It's what do we stack on top of that? That isn't how we get AGI. That's, I've not heard that before yet, that we don't, that doesn't fork off into AGI. I personally don't think so. Now, I might very well eat my words in, two years and when I'm serving my AI overlord. If it's two years in the future, comment underneath. I don't. So I think what this represents is a revolution in the way we search for information. Yeah, I imagine I'm not the only one here who chat GPT has replaced Google as my first go to when I have a a technical question. Yeah, I think it's safe to say it's eating the market share. Yeah. So I think that's where we're gonna see a lot of transformation. I think the amazing parts of the technology aren't that it could mimic this. Conscious being but the amazing part of the technology to me has always been it's pattern recognition Look, it's really hard to program something to like Yeah, take a picture of a hot dog and recognize that it's a hot dog. Yeah but chat GPT can do that Yeah, it's good visual. Yeah, these LLM's They can recognize things. These are things that are very hard to program for. And I'm like excited to see what kind of applications that we can get. To me the ideal use case for LLMs aren't Like a chatbot. I'm, I hate chatbots. I wish they would go away. I'd heard them described as the beginnings of the tiers of, not necessarily AGI, but the different types of interface, human machine interface that will be this LLM. Whether it's something that you audibly talk to, like an Alexa, or if it's something you actually visually interact with and show at things. Yeah. Whether it has a camera eye or whatever. Yeah. But yeah, like I can totally, I know enough about building things with code that's a hard thing to build. And there's this over here on the side, this emergent property of LLM. That can visually recognize and not only visually read a photograph and what's in it, but make inferences based on data that it's chewed up. Yeah. And I know, for example, I'm able to upload photos and tell me what typeface is this? Anything I want to know about the context of the picture. And it's able to, in detail, describe it to predictive text. Yeah. That's It's almost absurd. Yeah, it's almost absurd that it works, but it makes sense. How do you use AI on a daily basis? Yeah. So as a developer AI has been most useful for me if I'm working in a language that I'm not too familiar with. Simple syntax, that's really easy to look up. It's made that part of the job so much easier. I use it to automate, write little bash scripts that I otherwise don't want to do myself. But it's, saves me 10 minutes here, 15 minutes there, an hour maybe. All of those little, all of those little tasks. And then I just ask it questions, really. I still treat it as like a, and, I treat it almost like this assistant that can go look up information for me and then tell me you're paying for it. Yeah, exactly. Yeah. So like I have been really into investing lately or I'm trying to learn. I'm still very much a newbie, but I've been able to use chat GPT and have this tool that I can ask all my really dumb questions that I'm otherwise too embarrassed to ask someone who actually, is a proper adult. What, what's an ETF okay, stock prices, why do they move when the markets are closed? I simple things like that, I just don't understand, but I feel almost self conscious asking. Sure. Yeah. Yeah, I would too. Even though it's probably a comp, Most people don't know the answers to probably all of those at once, so it's okay to ask. But it's almost like having a conversation in your head or something. It's between you and OpenAI or whomever owns the conversation you're having. The conversational part of it, because we are such language driven creatures, typing it is a bit laborious, but once we can interact with voice, That's pretty compelling, I think. Yeah, it is. Soon enough. I think you touched on an interesting point, that it's a different way of interacting with computers. And I think in a way, it democratizes it because anyone can have a conversation with this thing. You don't need technical chops to be able to do it, and I'm in favor of that. Yeah, and it's not really of a judgmental mind, too, of your pedigree, or whatever you want to call it. I hope not, man. The questions that I ask it sometimes are just like, wow. Yeah, it What's interesting is the context of the chats too, that you can reference previous context and data that you've run through with it. I find that very useful on my day to day or week to week or however often I use it. Yeah, and agents will probably build on that. I imagine. Maybe even more useful. And having, it's just if you want anything done, period, very reductive. Break it down into its. It's separate pieces and then accomplish those in sort of order of operations. So if you can apply to a few of those separate pieces to an AI model and you can execute it just as if you did it, then it's very viable as something to speed up and squash costs and production costs and all of that. Something that I was able to ask it the other day was to cross reference and check itself. With more data, and to between itself determine, I think I was writing Photoshop scripts or something like that, similar to what you were talking about, like the The batch stuff that it can write for you, I think is JavaScript or something, along these lines, but That's the great part about it, is I can be a script kitty a little bit, Which is. Isn't that like somebody who just copies and pastes? Do I have that right? All, every, all devs copy and paste to some degree. Sure. It's like a, it's like part of our careers, but Yeah, it, I think it, it's very good at doing some of the more tedious coding tasks. Yeah, syntax, like you say, like checking for syntax. When I was first starting to learn coding just enough to be dangerous. The syntax was almost always where I went wrong or where I would throw an error or, something would fail. Yeah. Is a misplaced semicolon. So if you can have AI doing double duty while you're in the editor even is really cool. Not that doesn't already happen. We have autofill and like all of that good stuff that checks for you. But even pre run like pre running the code. That could be pretty cool. Yeah. Yeah, I think another area is just AI is also very good at explaining new things. If you ever touch on a technology that you're not familiar with, you can just ask, chat, chat GPT to explain this thing to me. It's like Stack Overflow Plus. Yeah. Sometimes. Yeah. Very cool. It's as if you could interact with conversations or threads within Stack Overflow and then ask it to actually, okay, do this and then basically spit out what you want to do. The result without you having to do all that within an editor. Yeah. Just for like listeners for context, Like stack overflow is this website we use. It just houses a ton of questions, and then other Developers will answer those questions. And, for a lot of our careers, when we come across An error in our code that we're not familiar with, Stack Overflow would often be like a first stop because chances are somebody else has already encountered that same problem. And thanks to the generosity of developers in general, someone who's overcome that problem usually answers it. The challenge before with Stack Overflow was that Your problem might still be unique in comparison to like what you find on Stack Overflow. So you have to translate that solution into your, to fit your problem exactly. And ChatGPT has been, is just way better at it because you can tell it your exact issue. It can try to resolve it for you. To the best of my knowledge it reads, it's able to read Stack Overflow, I think? It's probably trained on some Stack Overflow data. Yeah, that, that is a gray area at the moment, I think. Training material and, Yeah GPT 4 has already pretty much used up most of the internet. That's the scale of the training data. Yeah. And then what's next? 4. 0 or something? 4. 0 is like a miniature version. Yeah, 5. I just, I don't know. Again I'm a skeptic and I say that knowing full well that maybe, yeah. I appreciate a skeptic. I really do. Maybe it's, maybe it's my coping. I'm sure a lot of people are like, Oh, he's just coping. He's scared. He's going to lose his job. Maybe I'll eat those words, but I really just don't. I've been in this industry since 2016, and, which isn't a whole lot of time, there's people with way more experience than me, but I've already seen multiple hype cycles. Come and go. Yeah. Are you familiar with the gar, the Gartner Hype Cycle? No. Please. So it's just a simple model. It goes, okay. When a new technology comes, initially expectations are very high. Everyone's oh yes, this is gonna change the world. That is the peak of what are the, what is he call it, the peak of unrealistic expectations. And then there's the trough of disillusionment. The TR disillusionment. Yeah. That's when it comes back down again when people realize, oh, it's not. And then it'll slowly crawl back and that's where you achieve like enlightenment. Okay, this isn't going to solve all my problems for me, but it's a tool and it has its place. And hopefully that's where we'll end up with AI. But look, before AI came along, everyone was talking blockchain. Yeah. Blockchain this, blockchain that. Everything was going to be built on a blockchain. Web 3. 0 is here. Before blockchains, it was cryptocurrencies, right? Before cryptocurrencies, it was big data. Big data is going to transform your company by harnessing it. Yeah. I haven't seen any of those kind of, I guess reach the levels that they were touted on these like various, TED talks and et cetera. It's been a lot of boom and bust. It feels like in tech for the past five, five years or so, maybe even eight years. With not with a lot of payoff but not The hype cycle payoff like yeah top of the top. Yeah, and to be fair. I do think Like LLMs, I think generative AI is probably gonna transform our lives more so than say blockchains will But I don't see it being the thing that's gonna I don't see it being the scary monster that's gonna take everyone's jobs and We'll all just be you know, Living in some sort of subsidized. I like that skeptic perspective. That's actually a really positive one. Yeah. That's like a positive cynicism toward the. Yeah. Look, calculators didn't make everyone mathematicians. And then mobile phones and, everything else that's become. Like a part of our anatomy over time. Oh, yeah. As a digital device. Apps. Before big data, it was apps. Everything was going to be an app on your phone. Huh. Yeah. Huh. And these things happen in a way but not necessarily where they're the end all be all of the panacea to everything. For example, AI automating everything, and then universal basic income. That's the, a little bit probably too far into a utopian future, but it's probably somewhere happy in the middle to where, We have this renaissance of being able to produce way more than we've ever produced. Yeah, we are but we're producing a lot of What a sludge. Yeah, right a lot of a lot. We're slogging through a lot of Novelty, I guess is a positive way to put it a lot of new like just untold amounts of it that it's becoming hard to actually run the models because of all the All the quantity. Yeah Yeah, I think it's It's definitely been a boon for, scammers. Yeah. They have, they are now empowered to generate more content than ever before. Yeah. Aren't they always the early adopters, or almost always the scam? Crowd. Yeah, probably able to figure out how to head of the tech curve. Yeah, a little bit. Yeah. There's something to be said about like how people hack to gather things. How do you stay on top of trends? Tech trends in particular. For me, it's a lot of, so I have certain reddits, like subreddits like the rjangoreddit some like dev related ones that I subscribe to, and YouTube channels and TikTok, really. TikTok? Yeah. Tell me more about that. Yeah. When you run across something on TikTok, what what sort of information is it, and do you implement it in your day, in your work? No, a lot of times on TikTok, I just like to get other people's outlook on the market. Yeah. People who, have more experience than me, people who understand the business side of it, way better than me, so they can break things down, and then it's just my way of keeping my fingers on the pulse. Yeah, there's so many content creators who make it their job to present this information in such a bite sized way, short form, that's You can save for later. Very shareable. Yeah. That's super cool. And then YouTube old faithful. Yeah. A lot of like long form these video essays on YouTube that I really enjoy. A lot of my sort of views on AI are from watching content creators on social media. A lot of times they would come from a more computer science background. Maybe they were in the research, they're doing coding at a higher level than I am. Like they are, in academia. working on these sort of theory, the theory behind a lot of this, and it's from their opinions too that I think I derive my own, which is that if we're counting, if we're banking on LLMs solving everything and just being this, yeah, completely transformational technology I'm skeptical, and I think there's a lot of working professionals who are tech savvy and also skeptical. I think the people who are really hyping this up are, they have business interests. Investors. Yeah. Consumer investors. Yeah. And institutional, I'm sure. But from a, I really do love speaking about it from not so much an investment standpoint, but from a, like a practical producer, builder standpoint. Yeah, it just fascinates me because as a builder, as somebody who Pretty much my week to week is actually producing like a product So finding out ways to do that is doesn't harm the the quality or the hand touches Yeah so focusing more on the details that I think matter and then less on the kind of the white space in between the negative space of Building the scaffolding around what you want to do. Yeah, and the Foundational pieces Because to make anything, to build anything takes so much energy and effort that the amount that we can minimize that is really good. Really potent for tech in particular, I think. I don't really have anything else besides just to chat about AGI maybe, but is there anything that you have in mind, Derek, that you want to share? I think. Okay, so I think hopefully if anyone's listening to this and they're considering their own sort of career transition into tech, I can say that, the coding bootcamp was a great decision for me. It's probably the best like career move that I've ever made. Granted, this was in 2016. The market's a little different now. It's saturated with a lot of junior developers coming out of these boot camps. And. For any listeners out there who are considering it, I would still consider a boot camp. But what I've learned on my journey was that it, number one, be prepared to really hustle for that first job. That first job is going to be the hardest one, probably in your entire career to land. That first technical role I've had friends who were unemployed for probably almost a year. Before they found this post bootcamp, they were unemployed for almost a year before they found a gig. However, now eight years on, I would say outcomes are like 90 percent good just based on my own sort of anecdotal experience. Everyone's doing pretty well. Everyone settled into their careers. It yeah, it was really, it's, this is a skill set that's very empowering. And I think as Society, as we continue, I still see this shortage in technical skills and there's a lot of tech roles I think are still going to be difficult to fill moving forward. So there's still very much an opportunity. It's just that You have to figure out how to set yourself apart when you come out of these boot camps because they're going to make you have all the same resume with your three projects. And it's even the formatting is similar and I say set yourself apart somehow. If you're if other people are putting links to the GitHub repo to their projects. Go a step further. Deploy it on a free hosting service like Heroku, or even AWS is great at giving you a lot of free hours and compute time. Deploy it so that when you send your resume to recruiters and people who are interviewing, they are much more likely to click on your project. No one's going to go look at your code, pull it down. I've sat in on interviews. Never once did we ever go into the repo and actually look at the physical code. Go line by line. No, we're not doing that. But if you have it, deployed somewhere, great. If you can't do that, capture a GIF of it running on your local machine and then send that. Just anything to make the interviewer's job easier. And then as far as, Most bootcamps are full stack, so full stack being front end and back end training. But I think, I'm a back end developer out of a bootcamp because I really, I truly enjoy the back end. But if you don't have a preference, it seems to me that it's a safer bet and you'll have a smoother time going more like going more front end. Because I think the front end is a little bit more encapsulated. You only have to know, CSS, JavaScript and HTML, right? JavaScript and TypeScript and then various like JavaScript frameworks. It gets deep. It gets very deep on the front end as well, but you're not having to deal with so much like legacy code that's in the back end. Like you never know what you're going to run into on the back end. So yeah, if I can have a, if I can give a little bit of advice. Yeah. So that functional code that's actually deployed out there like in a web app or through Heroku or some sort of method where they could play around with the functionality of what you've built. That's a really, yeah, that's really great advice. Yeah. So I think that's great. Just, yeah. Set yourself apart a little bit. Be patient. But I think ultimately I do believe that The market will bounce back. The job market, I know, is atrocious right now, especially in tech. Also consider, you don't have to work in tech. I'm a software developer, but I don't, I would consider myself more tech adjacent right now, because I work for, this creative agency. We're not really some of our clients are, tech companies, or we use a lot of the same technologies, but technically we are Like in advertising, and I think that's insulated me a little bit from all the tech layoffs. Yeah, like we're living through the disillusionments right now with these like tech jobs, right? For the past 10, 15 years, Google, Facebook, Amazon, all your fang companies, they were like dream jobs, right? But now people are starting to see with those dream jobs come the golden handcuffs and then the crazy hours. And most importantly, I think psychologically this, you never have a feeling of security. You can be laid off at any minute, I think that's a tough thing to live with sometimes. So consider that. You don't have to just go in tech. Every field needs web developers. Now, every field needs programmers. Huh? You don't even have to be a programmer coming out of the. Boot camps if you don't, or web developer if you don't really want to. I've known people who transitioned into embedded systems Programming, some people just taught themselves C and wanted to go in a different direction, didn't Want to make web applications all day. That's fair. I don't think there's a window into those types of careers, technical or otherwise. No. Any like project managers or any Yes, absolutely. Those come out of the bootcamps as well. Cool. Some people didn't wanna just be a pure engineer, wanted to be more of a technical project manager. Someone who is more, coordinating the project together and then using their technical expertise to make sure everyone is on top of, making sure that we're building, we're a team of developers and designers and all that, that we're all building towards the same goal, and that's whatever our client needs. Yeah. Yeah. Right on. No, thank you, Derek. Yeah, no problem. Thanks for having me.