Episode #114: Serverless for Salary Transparency with Kesha Williams

October 11, 2021 • 45 minutes

On this episode, Rebecca and Jeremy chat with Kesha Williams about her 26 year journey in tech, how the cloud enabled her path to becoming an AWS Machine Learning Hero, her motivations behind building Salary Overflow, how serverless made it easier, and much more.

Kesha Williams is an award-winning software engineer and technology leader teaching others how to transform their lives through technology. Forbes, Amazon Web Services (AWS), and Oracle have applauded her contributions to the technology community, and she has spoken on the TED stage about the transformative power of artificial intelligence (AI). Amazon recognized her pioneering work in AI with both its AWS Machine Learning Hero and Alexa Champion honors — the first person to receive both. Williams was named Mentor of the Year by Women Tech Network and received the Innovator Award from Hospitality Technology. She has launched several successful startups and appears in the 2020 tech documentary, "Hello World: The Film." Additionally, Williams serves on the Board of Directors for Women in Voice and as a mentor to women in tech. 


Rebecca: Hey, everyone. I'm Rebecca Marshburn.

Jeremy: And I'm Jeremy Daly.

Rebecca: And you're listening to Serverless Chats.

Jeremy: Hey, Rebecca. How are you doing? And I'm going to stop you before you start telling me about your weekend, be`cause I know I'm always asking you how you're doing on your weekend. But I actually was curious if you had binged any shows lately or watched any interesting movies.

Rebecca: That's a great question. I have not really binged any shows. I'm much more of a lay on the couch and stare at the wall and listen to the same record over and over and over. It was actually something that really made my roommates in college mad. They'd be like, "Please just switch the side of the record."

Jeremy: It's probably better for your mental health [inaudible 00:00:42].

Rebecca: But do you have anything that you recommend I should binge? Should I be in the market?

Jeremy: Well, as you know, I am a soccer coach, because I have two teenage daughters that play soccer. So, the show that I've been watching lately has been Ted Lasso on Apple Plus, which is an absolutely amazing show. And it's more than just about soccer. It's feel good, but also delves into mental health and some of these other things. So, anyways, really interesting show, and super funny. Jason Sudeikis is hilarious in it. So, anyways, that's something that I would recommend.

Rebecca: All right, all right. Well, I'm wondering if the guest I'm about to introduce, who is a superhuman mother of three, I'm wondering if she has any similar experiences in terms of sporting events, and then how that informs the types of shows that she likes to watch. But first, before we get to that question, I will introduce her. Our guest this week is an award winning software engineer, an AWS machine learning hero, a principal training architect for AWS at A Cloud Guru, which is now a Pluralsight company, super exciting to see that transformation, and the creator of Salary Overflow, Kesha Williams. Hey, Kesha. Thank you so much for joining us.

Kesha: Hi, Rebecca and Jeremy. I'm super happy to be here.

Rebecca: Hey.

Jeremy: Now, as a mother of three, do you have any time to watch TV?

Kesha: Well, my kids are... They're getting older now, so I do actually have time. But I remember those days, Jeremy, when all three of them played soccer, and my Saturdays were crazy.

Jeremy: What's a Saturday, right?

Kesha: Yeah, exactly. From one soccer game to the next soccer game on this side of town to the next side of town. It was fun. I do miss those days.

Jeremy: Yeah, I am trying to enjoy it, because I know that soon it's going to be... They're going to be off somewhere else, and I'm going to have nothing to do on Saturday. So, I am enjoying it while it lasts.

Kesha: That's good.

Rebecca: So, to kick this off, and to kick it off all the way back to the idea of childhood and remembering those days, Kesha, when talking about your own childhood, you said in one hand, you had a Barbie doll, and the other hand, you had a computer. And it seems like that led to an amazing journey in tech. You've had 26 years in tech, and you've worked for the NSA, and then you eventually moved to Cloud. Can you give us a little bit of your background for context for our listeners?

Kesha: Sure. So, I've been in IT for 26 years. I honestly do not know where all of the time went, because it's just been such a fun and exciting journey. And I really built the bulk of my career in the Java software engineering space. So, this will date me, but I remember when Java came out, because at the time, I was doing C++. And I heard about this new language, and I researched it. And when I learned about it, I thought, this language is going to take over the world. I don't want to do C++ anymore. And so, I actually left the organization where I was working at, and I moved to Delta Airlines so that I could learn Java on the job. And I did that for a very, very long time. And throughout that time of doing Java web development, I held several titles like web developer, software engineer, technical lead, and then I eventually moved into a leadership position managing a team of engineers.

Kesha: And I would say about seven or eight years ago, I started to lose track of the time, maybe seven or eight years ago, I was introduced to the cloud. And I really felt about the cloud how I felt about Java way back then. I'm like, okay, this is going to revolutionize how we deliver systems. And so, I fell in love, really, with AWS. And then maybe three years ago, I started playing around with machine learning on AWS. And yeah, the rest is history.

Jeremy: Well, I think with the pandemic, we can add or subtract a year, because we always forget 2020 even exists, or hopefully we can forget 2020 existed, for some things anyways. But so, you had, obviously, a very long career in tech. You've done a bunch different things like you said, moving from C++ to Java, then getting into the cloud. And there's just this whole array of different roles that you've held, whether it's in management positions, or just being an engineer at some of these places. And I'm really curious, because everyone's journey in tech is different. And we hear a lot of people who are like, as you said, you started with a computer in one hand, a Barbie doll in the other. So, you were into tech very young. A lot of people change later on in their careers.

Jeremy: But one of the things that I think we know about tech, and this is just, unfortunately, one of these horrible labels that tech gets, is just the diversity in tech is terrible, right? And that we've got salary discrepancies. We get people treated differently. There's people, you'll start at a company, and then have a horrible experience, because they're not a brogrammer, right? They have those experiences. So, you created this thing called salaryoverflow.com. And I'm really curious, just, what were those experiences that you had that led you to creating this, and what was the need for it?

Kesha: Well, in tech, we always hear about this gender pay gap. And being an African American woman working in tech, I firsthand have experienced that. So, my very first experience with it was when I actually graduated from college, and I was working my first job. And I remember, I believe it was the first week, all of the new hires, we were sitting around the lunchroom table, just chit chatting. And then somehow the topic about salary, starting salaries came up. And as everyone went around the table discussing their starting salary, I realized that I was making $20,000 less than everyone else. And so, I was the only woman and the only person of color. And so, for me, that was very eye opening. And then as I just moved throughout my career, and just throughout my journey, I would always notice that my colleagues, they were able to take all of these expensive lavish vacations. And for me, I was a single mom, with a mother of three, but I was debt free, and so I managed my money very well. And I was never able to fly my kids off to Hawaii. And so, that just made me think to myself, okay, they must make a lot more money than I do if they can afford those expensive vacations.

Kesha: And so, because of my firsthand experiences, I wanted to do something to bring more transparency to salaries in tech, to hopefully just reduce that gender pay gap. And that's why I created... Well, that's one of the reasons why I created Salary Overflow. I also just love working on side projects to keep my skills current. And so, I wanted to play around with Amplify and some of the services on AWS that I'd never used before. So, I think Salary Overflow was a win-win.

Rebecca: Okay, so there are so many underlying foundational themes that I am really eager to address there. But before we get too far into that, let's talk a bit about, you went from Java, and then you found, or you're in... You write... You prefer Java, but then you found the AWS Cloud, right, and you're like, this is going to transform the way we work. And then within the last seven to eight years, that also has... You have become a machine learning hero and an Alexa champion, and you've become a voice not only for teaching others how to use these successfully, but also as an advocate to say, hey, this is... As a African American woman of color, there's not enough diversity. There should be more than just me as the face of this. Before we go too far into some of those other media topics, I'd love to dive into, we know why you started in Java. What prompted you to start with cloud? What made you think that that cloud was the future?

Kesha: Well, for me, as I've always considered myself a software engineer, developing and building, bringing ideas to life. When I was introduced to AWS, I realized that I didn't need a team to build or bring these ideas to life. I could use AWS to quickly, literally in a matter of weeks, build an application that normally it would take a team of five developers to build over several months. So, that's really what got me excited about the cloud. Just for me, my ability to quickly innovate and deploy solutions to production, it's just... It's amazing. That's why I love AWS.

Jeremy: So, again, the idea of AWS being this vehicle for small teams or single developers to be able to build these massive things is, I think, interesting to the way that AWS has brought machine learning to the forefront with all the SageMaker tools and all the different things. So, maybe you can expand a little bit on that, because I know you've only been doing machine learning for maybe what, three or four years, which is three or four more years than I've been really doing it. So, I basically use the API's. That's about the extent that I go with machine learning. But what was it about the openness or the availability of tools or software that even let you get into machine learning? Because I'm thinking, normally, you'd be running massive EC2 instances and things like that. Just what are some of those things that made it easy for you to explore machine learning?

Kesha: Right, and I always tell people, if it weren't for AWS, I would have never gotten into machine learning. So, when I initially started wanting to learn about machine learning, again, I always find excitement in these technologies. So, first, it was Java, then it was cloud, and then machine learning. And again, I just feel like machine learning is this disruptive technology that is just going to continue to impact our lives on a daily basis. And so, I wanted to learn more about it. And when I first started my journey, of course, I went to AWS, and there was a service called AWS Machine Learning. Nice name. And that service is no longer available to new developers. But for me, at the time, I was just starting to explore machine learning. And it really abstracted away a lot of the deep technical details. And it just gave me an easy way to get started, and it made this topic very approachable. So, I wasn't afraid, because before, I always felt like you had to be a research scientist working in a lab somewhere in order to even understand and use machine learning. But with that AWS Machine Learning service, it was a great introduction.

Kesha: And there, once I became comfortable, I understood the machine learning lifecycle, I understood just all of the machine learning concepts, then I started to explore and dig deeper. And that's when I started learning using SageMaker. And then SageMaker just allowed me to continue to peel back those layers and go deeper and deeper and deeper until I was actually ready to start writing my own custom training code. And so, I think for me, that was the great benefit of AWS. I was able to start at this very high level and just start peeling back the layers to get really deep and really technical. And it was something that I could do on my own time. And it just made it very easy. Like I said, if it weren't for AWS, I would not be a machine learning hero. I would not have worked with machine learning.

Jeremy: And you don't need to be a rocket scientist, right? You don't need to be a math... You don't need a PhD in math to do some of this stuff. I remember very early on trying to play around with TensorFlow, or what... And I was just like, oh, my God, eighth grade algebra, I can help my kid with her homework, but once I get past that, I really got to start thinking about it.

Kesha: Yeah, and I will say I did have to learn Python. And to this day, I still want to put a semi-colon on the end of all of my Python lines of code. And I still don't get this indentation, why?

Jeremy: Oh, yeah. No, it's like [inaudible 00:13:24], right? Indentation is important.

Kesha: Why?

Rebecca: So, I was talking to Jeremy before, obviously, before we were interviewing you right now, and I was like, oh, man, I got so into watching Kesha's videos, and so much of her content that I almost forgot to come back and form questions, because I just was so... I was getting sucked into all of them. I love doing research on guests, but yours were particularly compelling. And one of the talks that I really liked is the one you gave in 2019 at re:Invent, Future Proof Your Career: Java Developer to Machine Learning Practitioner. And in it, you really lay out the levels of adoption, which it sounds like you just walked us through your own path. But the levels, quickly, that you lay out our level one, you're using recognition, and poly, and lax, and you're using ML models created by others. And then level two, you make your own models, which was AWS machine learning, and now it's basically been crafted into SageMaker. And then there's level three, where you use toolkits and interfaces and frameworks to write your own learning algorithms. I'm wondering if you could talk a little bit about... A little bit more perhaps about the leaps that maybe you needed to make personally or that you see others need to make in order to up level.

Kesha: Right, so if you're trying to start with machine learning, definitely learn the concepts. So, the very high level concepts, like what is a model? When someone says they need to train a model, what does that mean? And so, once you learn, I call it the machine learning lifecycle, then you're able to actually implement that lifecycle using tools. And so, that's how I always tell people to get started. First, start with just the very high level concepts, and then once you're good with those concepts, just continue to peel back the layers and go deeper and deeper and deeper. And like I said, all of the tools and services on AWS allow you to be as high level or as detailed as you want to be.

Rebecca: This is a question that I think relates maybe back to Salary Overflow, but this idea of what it means to be a minority in tech. And something with machine learning, right, that we hear so often, or science and technology studies more broadly, is if you put that information in, you're going to get bad information out. And so, there's this, when you're at that level three, and you're using, you're writing your own machine learning algorithms, how do you create algorithms that have as little bias, ideal state, no bias as possible? And I'm wondering as being both a woman and a minority, an African American woman in tech, if you approach writing your algorithms perhaps differently than you've seen others approach it, or if you've found different ways that your algorithms maybe come out a little differently because of your previous experience, because of your life experience.

Kesha: Right, so that's something to me that's very important when we talk about the bias in machine learning. So, I can tell you about the very first machine learning model that I developed. It was used to predict crime. And the data that I've found that was freely available was UK crime data. And one of the data points in that was race. And so, when I was building my machine learning model to predict crime, because I'm an African American, I personally decided to take race out of the equation. Now, someone else may not have considered doing that if they've had a different experience. And so, I think whenever we're building machine learning models, you really need to have a diverse team so that certain things are considered upfront when we look at the data, when we test the model. Because we've heard and seen several horror stories about computer vision, anything related to machine learning having a bias. So yeah, it's very important, especially how we utilize machine learning today in our lives. It's very important to build models and solutions without bias.

Jeremy: Yeah.

Rebecca: Do you see anyone doing it well today, or do you see perhaps companies or individuals who are teachers and educators around building algorithms that are setting up much better guardrails or criteria to be like, hey, this is actually going to get you an algorithm that's going to return something that has a ton of bias in it? Are there any great resources that you would recommend in terms of building better unbiased algorithms?

Kesha: Well, I can tell you, I jumped for joy, literally jumped for joy when SageMaker Clarify was released at re:Invent last year. And I was just like, yes, this is a great first step to make sure the models that we're building are bias free. And so, I was just really happy to see AWS make that a service, a part of SageMaker, and just making it available, and just even recognizing that it's an issue, and it's something that we need to address. So, I would say definitely, if you're building machine learning solutions on AWS, incorporate SageMaker Clarify into your solution.

Jeremy: Yeah, that's amazing. So, there's clearly bias. We know this in machine learning. A lot of these, you mentioned that the visualization stuff, not being able to detect different skin colors with facial recognition and things like that. And again, part of it is because the test sets, right, who's actually building it, who's testing it, all of that bias comes through. And I don't think all of it's intentional. I think some of it, it just happens because people don't think about it the way that somebody with different experiences might think about it. And I want to take that... I want to pull the thread a little bit on the idea of bias, and you mentioned this early on, this idea of you making $20,000 less than everybody else sitting at that table. Those biases come in, again, whether they're intentional or non-intentional, whatever, the point is, is it exists, right? So, regardless of how it happens, it's there. It's systematic, and it's just in the system now, where no matter what you do, you're going to see those biases come through when it comes to salary data. So, let's dive more into Salary Overflow a little bit. You mentioned the reasons why you wanted to build it, but let's actually talk about what it actually is and what it actually does.

Kesha: Sure, so Salary Overflow is a web application right now. It's in the initial, I call it the data collection phase, where you're able to go in and enter your salary information, and then search to see the additional salaries within the system. And so, it's broken down by location, job title, and the salary amount being paid. And so, that's the first phase, I call it the data collection phase. And honestly, I put it out there just to see. It was more of the MVP, and the response that I received was very overwhelming. And so, it made me realize that yes, this is something that's very important. And I even had a company reach out to try to buy the IP, but I decided to hold on to it, because I have just so many, just amazing and just exciting plans for the next phase of the application.

Kesha: Definitely, I want to incorporate machine learning to make salary predictions. I want to incorporate a salary negotiation piece to help people better negotiate salaries. I want to have reports. So, I imagine there's a woman out there, she has this job offer in hand, and she really wants to know if she's being paid fairly. She can go into Salary Overflow, print a report for that same job title, same location, same years of experience, and say, hey, look, you are paying me $10,000 under the market, and it would help her negotiate her salary. So, I just have a lot of... I just get really excited when I think about the future of Salary Overflow and how it can help people.

Jeremy: It would be interesting to have a feedback loop in there as well to say if somebody did get a job offer that was below market rate, or whatever, to report that back into the system that you got an offer from this specific company or whatever, and you could do some of that as well.

Kesha: Yes, that is really cool.

Jeremy: And you can maybe even rate companies on their transparency level-

Kesha: I need to get a notepad. I like that. I like that idea too.

Jeremy: Write it down, write it down. So, your reason for doing this clearly is amazing. And this is just really, really needs to be done. But we're at Serverless Podcast, so I would be remiss if I didn't ask about the architecture that you're using, because you built this using all serverless stuff, right?

Kesha: Yes, and I really relied on AWS Amplify. So, I've wanted to experiment with this full stack development, web application development tool that came out. And so, with AWS Amplify, you're able to really speed up the delivery of a web application. And so, the front end of the application currently is served up through S3 and CloudFront. And then the user authentication, I use Cognito. And then between the database and the front end, I'm using AppSync. That was my very first time using AppSync and GraphQL. I'm usually a Rest person. And I found GraphQL to be very flexible. And then on the back end, I'm using Aurora Serverless. And so, there's also a Lambda function in there as well whenever there's a new user that signs up. I use a Lambda function to create a record in that database. But yeah, just the overall experience and flow. And also, this is my first time using React. So, I use React on the front end. I'm typically an Angular person. So, it was just-

Jeremy: We won't hold that against you.

Kesha: It was just a great overall experience, and I was able to learn a lot of new tools and technologies, so yeah.

Rebecca: So, I'm wondering with that, have any of those services catapulted into being your favorite AWS services? Or is there anything where you're like, hey, I'm actually not going to build an app without Amplify again because it is this full stack experience and it enabled me to learn these things-

Kesha: Oh, yeah.

Rebecca: ...interact with these services. You don't have to know too much. I shouldn't say know too much, but you can start from the starting point.

Kesha: Right, so I really love Amplify. I'm a huge Amplify fan. And there's even a component in there. There's the automatic CICD. So, if you hook it into your version control system as you merge changes to, let's say, so I have like a dev test and prod branch, as I'm merge changes, it automatically deploys my code to production. And so, that was really built in to Amplify. And so, it just gives you this whole experience for developing and deploying applications I call the right way. And so, I am definitely an AWS Amplify fan, a huge fan.

Rebecca: I know some teammates on the Amplify team, and they're like, yes-

Kesha: [crosstalk 00:25:33], I love it.

Rebecca: Speak it.

Kesha: Team of one.

Jeremy: We just had a conversation with Ali Spittel, who's on the-

Kesha: Oh, really?

Jeremy: She's the developer advocate for... The lead developer advocate for that team. And yeah, we just had a good... We just had a great episode with her as well. So, you mentioned speed a number of times, just gets you where you're going fast. Clearly launching EC2 instances, or even setting up containers and doing some of the other things, just how much productivity do you think you gained from using a serverless approach? And clearly, Amplify is serverless approach on steroids, but what type of productivity gains do you think there are here for developers that are traditional EC2, VM people, or container people moving to something more serverless?

Kesha: Well, I can tell you with the Salary Overflow application, the MVP, I was able to have that finished in a matter of weeks. Literally a matter of weeks, it was in production. Not even two weeks. That's why I quickly did it. I put it out there to see if anyone would find it valuable. And now, when I look at the UI, so I told you, it was my very first time using React. And so, it's a very basic UI. And now, I'm like, okay, I can really take this to the next level, because people actually like it.

Rebecca: Yeah, and for those who haven't experienced Salary Overflow, even if it's a basic UI, you go in, and you have to create your user first before you see some of this data. And you get an SMS message, right, with a passcode, and then you enter the passcode, and then it returns you to the sign in. And then now you have your username created, and your password, and then you get in, and you get to start seeing this data. But while it may seem basic to someone who's a programmer, I think for a lot of people, you're like, that's actually a considerable amount of work to make that flow happen so seamlessly-

Kesha: That's true.

Rebecca: And to onboard someone onto some sort of product. And so, I do think that you probably are not even giving yourself enough credit in terms of being able to set that up in a matter of weeks. And that's only the onboarding flow, right? Before you even get into the data and the reports and ingesting all this data and all of that. So, I highly recommend people to check it out.

Jeremy: Yeah, and also, I just say too, I have built a lot of React or Vue apps as well, and I am not a designer, so they always look clunky and basic and whatever. And it always irks me a little bit when that designer lays this nice little skin over, and you're like, how did you make it look so good? So-

Kesha: Oh, I know that feeling.

Jeremy: You're not a designer like me. Anything that looks basic, there's a lot happening back there. I can attest to that. But yeah, it doesn't always look as great as we want it to, but props to designers who can do that and make things look so beautiful.

Kesha: Definitely.

Rebecca: Yeah, and I wanted to dive in a little bit more. I know that Amplify as a full stack also has databases that it's going to already use, and it's using Aurora. And I was wondering though, when we go back to machine learning in general and the amount of data that you have to ingest, and then it's not only that, right, it's ingesting it, and then you have to clean it, and format it, transform it, you have to check it. I'm wondering if outside of Amplify and what would be coming in that full stack experience, if there are databases that you love to use for machine learning data in general, and what you see in terms of pros, cons, benefits, advantages, disadvantages to different databases for performing all of that on your data.

Kesha: So, specifically for machine learning, I use S3. So, I don't really use an official database. So, S3, when you think about SageMaker, SageMaker just integrates really natively with S3. And so, that's how I store, for machine learning, that's how I store all of my data.

Rebecca: I love that. Classic.

Jeremy: And S3, by the way, it's kind of its own database. People don't think of it as a database, but it's like a key value store that can just store massive amounts of data. And also, with S3 Select and some of those other things, you can actually run queries on it. So, it's not as good as a database as Amazon Route 53 is, but it's still, as Cory Crean would say. So, let's talk about entrepreneurship a little bit, because you mentioned you had an offer to buy Salary Overflow. And I'm just curious in terms of, for other people that are building, and you mentioned the speed is there, and you don't need a bunch of people to build this. It can be a single person that can actually get something up and running really fast.

Jeremy: But for those people who are looking to do this, because this... The cloud enables so much. It's amazing what the cloud enables you to do now, especially for the small teams, like you said, especially for the kid who wants to get on the free tier, and sitting in their dorm room somewhere, and she can build a whole app or whatever, and it costs $10 maybe to run it. So, just in terms of people who are looking, or some advice maybe, some thoughts and recommendations, some learnings, whatever, that as you've gone through this process, what would you want to say to people who are thinking about starting on AWS and building a startup idea?

Kesha: I would say definitely start with just understanding all of the services out there and what's available to you. You don't have to go deep, deep, but just at a high level, know what's available. And so, I always direct people to start with the certified cloud practitioner, that level of knowledge. Especially as an entrepreneur, I would recommend that you not always rely on a technical person to just come in and work their magic, and then just leave you with everything. Just really try to understand just the basics of the technology, what's available, and then go from there. So yeah, I would definitely recommend with just understanding the basics of AWS, and then you can start to architect your system and work with solutions architects to help you architect your system. And when they talk all of their tech jargon, you'll understand what they're talking about, and they'll think you're really smart.

Jeremy: And there's probably a few courses they can take over at A Cloud Guru that would help you out with that?

Kesha: Oh, yes, definitely. Definitely, you can take my certified solutions... No, my CCP course. You can take that course. And it's a great, fun introduction to AWS.

Rebecca: When I first started working at AWS, where I'm no longer there now, but I was there for about four years. And so, I guess I started in 2016. And I still have my book when I took A Cloud Guru's introduction to AWS, or the 12 to 15 hour course. And then I was looking at the Lambda and serverless and some of the other specific courses.

Kesha: Nice.

Rebecca: But I still have my notebook that has this menu. I guess listeners can't see, I'm holding my hands really far away from each other. It's a really thick book of notes. And I did find it so helpful to even introduce me to the vocabulary that I didn't know I didn't know to start having those conversations.

Kesha: Right. That's awesome.

Rebecca: Yeah, it was... It is, it is, definitely. Yeah, so Kesha, I really admire what you're doing with Salary Overflow. And when I opened my email this morning, I thought it was very apropos that this person, Lenny Rachitsky, who has Lenny's Newsletter, and it's a newsletter that basically offers advice to people in the tech world and the PM world. Anyway, his edition today featured Niya Dragova. I'm not sure if I'm saying that correctly. But it's the co-founder of Candor. And the whole theme around the newsletter was the 10 commandments of salary negotiation.

Kesha: Oh, you have to forward that to me.

Rebecca: I will. I will absolutely forward it to you. And it's pretty cool. It has 10 steps around, hey, this is what you're going to experience across all these moments, and negotiation actually starts earlier than you think. And here's what the recruiter is going to ask you around, let's say, before they start conversations with a team, right, they say... Or before they start interviews, the recruiter might say, "Okay, what are you looking for as a number?" And then Niya recommends, hey, you don't give a number. You say I'd love to meet the team. I would love to know what your salary band is. So, it's more about how to diplomatically but advocate for yourself even early in the process before you try to start advocating for yourself what is actually step eight out of 10. So, I'll definitely forward that to you.

Rebecca: And I found it just so lovely that I opened that this morning when we were going to be having this Salary Overflow talk today with you. And I'm wondering if there are moments or places where you yourself go for learning and educational resources around salary best practices, negotiation best practices, or other ways that you look for things to influence or shape the way that you build your own app, or anything you could recommend to folks around. Where should I go to learn more about how to advocate for myself when it comes to compensation?

Kesha: Well, for me, it's really built off of, I guess, my 26 years of experience in tech and interviewing, and just being in the industry for so long. But I always recommend to people, and I still do this now, even if you're not looking for a job, but you know maybe in five years, I'd like to be a senior software engineer, I often just look at the current job openings out there, the job postings, to see what the salary range is. Not every single job will tell you what the range is, but a lot of them do. So, I think it's very important for you to know your industry, and know what the different job titles pay, and job postings is a great way to do that. And of course, Salary Overflow.

Rebecca: It's an interesting approach that I... It's working... It's essentially working backwards, right, from where you think you might want to go, and then understanding all of the goalposts or milestones along the way, rather than learning them once you get there. That's a really great idea, yeah.

Jeremy: I'm a little bit curious too how, and maybe this is some data that you've been collecting as well, but how geography plays into salary as well. It used to be very much so like, if you lived in Boston, and you worked in Boston, and you lived in San Francisco, and you worked in San Francisco, you could have much higher salaries than if you worked in some suburb of Columbus or something like that. And I'm curious if, especially with the world shifting to so much remote, you're now able to hire somebody from Oklahoma who is a brilliant programmer, and he or she has a much lower cost of living, but how does that affect salaries? And are you seeing any companies, and maybe this is just... I don't know if you've got this data, if you've dug into it yet, but is that something, one, have you seen companies paying equitable wages across geographies, or do you think they should?

Kesha: I definitely think you should be paid based on the work that you're doing. And so, I remember when this pandemic started, and everybody was working from home. And then some people took that as an opportunity, let's say they lived in San Francisco, and now they wanted to move to Florida, and some companies saying, "Oh, you're moving to Florida? Well, guess what? You're going to get paid less." I don't think that's the right approach. Now, I do believe that I'm going to start seeing a change in the salary data that's being entered into Salary Overflow, because a lot of people are starting to work from home now because of the pandemic. Right now, the data is still showing, I guess, I call it the old way, the pre pandemic world, where if you work in California, you're paid more than, let's say, you work in Florida. I don't know why I keep pointing to Florida. I live in Georgia. Maybe I want to visit Florida. But I do think we're going to start seeing this shift in the salary data. But I want to see it shift to where you're paid for the job that you're doing.

Jeremy: Right. Yeah, people should be paid what they're worth, right? That's always one of those things where it always bothered me when it's like... Especially overseas, when you have people say, "Oh, well, we can get these programmers for $12 an hour," or something like that. I was always like, it doesn't seem like that's the approach that you want to take when you're hiring people.

Kesha: Yeah.

Rebecca: I'm curious too with your work on Salary Overflow. You said that someone, right, wanted to buy your IP. I can only imagine that while that's one instance, I'm sure a lot of companies, or I guess I hope a lot of companies, are at least contacting you to better understand their own practices perhaps. Or do you have people reaching out in terms of advice or consulting or just taking a hard look at themselves? Or is that something that you want to do if that's not happening right now?

Kesha: Yeah, that's definitely one of the future phases of Salary Overflow, where corporations can see or visualize their own data in the system. And if they want to ask the question, well, how well do we pay African American women coders, or how can we compare that to what we pay male coders? So, that is definitely a goal and a dream of mine, for organizations to be able to visualize their own data and answer those questions using Salary Overflow.

Jeremy: Well, that's amazing. Well, Kesha, listen, this has been great. This is awesome stuff that you're working on. I hope all of these future enhancements, maybe not the ML stuff, but everything else will continue to use serverless.

Kesha: Oh, of course.

Jeremy: If people want to get a hold of you on Twitter or anywhere else, or learn more about the courses you teach for A Cloud Guru, check out Stack Overflow... Check out Salary Overflow, how do they do that?

Kesha: And before I answer that, the funny thing is, that's why I named Salary Overflow. Because when you think about Stack Overflow, now, when you switch that to Salary Overflow, it's the one place you go when you have questions about your salary. So yeah.

Jeremy: Brilliant, brilliant.

Rebecca: It is brilliant. I actually caught that. I was like, I think I know the nod she's making. It's a great nod.

Kesha: Yep. And you can find me on social media. On Twitter, it's Kesha, K-E-S-H-A, Willz, W-I-L-L-Z, or Z if you're in the UK. And LinkedIn, it's... I always fuss at myself for making this my LinkedIn, but it's Java Rockstar Kesha.

Jeremy: I think it's a perfect title.

Rebecca: Yes.

Jeremy: I think it's perfect.

Kesha: That was a long time ago.

Rebecca: [Stiffergirl87 00:41:23] over here.

Jeremy: See, I would have chosen a name like that, and then showed my kids just as a dad joke. I would have been like, yep.

Kesha: Yeah, and I feel like it's too late to change it now, because that link is everywhere.

Jeremy: Everywhere, right.

Kesha: Everywhere.

Rebecca: Kesha, that's going to be an NFT not too long from now, it's the screenshot of the URL of Kesha's LinkedIn. That's excellent. You are just future proofing your career again right now.

Jeremy: And then Salary Overflow is just salaryoverflow.com. And then kesha.tech is your other blog there. And we will put all of this stuff in the show notes too so listeners can go and find links to all the amazing things that you have done. Thanks again for being here.

Rebecca: Yeah, thank you so much.

Kesha: My pleasure. My pleasure. It was a lot of fun.