Tech Hiring in 2023: Trends, Predictions & Strategies for Success | Datapeople's Maryam Jahanshahi

0:00the U.S. I think spends 17 billion dollars on executive search every year. and it's not just like senior level roles where that's happening we see that's happened in junior level software engineer roles and i think what i think is a really exciting trend is unfortunately caused by the economic situation we need to like people need to be able to talk about efficiencies in the process and that's where their focus is and so that's causing recruiting to sort of be like what are the costs of all of our channels and which ones are actually doing well for us it's becoming a lot more data-driven than it's been in the past.

0:36At DevInrupted, we work to give engineering leaders actionable ways to improve their teams. That's why we're producing a three-part summer workshop series with Linear B. Each of the three workshops will explore the processes that elite software engineering organizations and executives use to deliver better business outcomes and reduce cycle time by 47 % on average in just 120 days. You'll learn how to assess your current performance and benchmark it against industry averages, streamline processes through automation, and improve business outcomes through resource allocation. Learn from the best and take your team to the next level.

1:10Visit our website to learn more and secure your spot today. We are back on Dev Interrupted. I'm your host, Conor Bronson, and we're live from New York with another incredible guest. Welcome to the show, Miriam Jahansahi. Thanks, Conor. I'm excited to be here. And I really love that you're here because we don't talk to data scientists that much. And you are not only the head of R &D and a data scientist at Data People, you are also a co-founder of that company. I am a co-founder. I also work very strongly with engineers. And so, like, I'm always up in their code and in their pull requests and all the fun side of things.

1:47I guess that's one of the things when you get to be a technical co-founder, you sort of have to run the gamut of all the different things that you do. Yeah, it was fun talking to you as we were kind of getting set up. And you mentioned you kind of had this opposite journey where you really dove into the data side. They were becoming like this strong data scientist. And then you realized you wanted to add these data engineering skills to the table. Yeah, it's an unusual experience. I think part of the reason why I had to do it was I had to figure out the systems that we needed to analyze data to kind of get a data-driven product.

2:17And so my role now is sort of, it is such a weird mishmash. I was talking to my co-founder about it the other day. I'm neither like, nor do I run engineering, nor do I run like the data side of things, but it's sort of almost technical product management. You're the fusion between the two of them. It's like a weird mix of many different things. And so we're realizing that, you know, that requires a certain level of skills and different types of agility. And so it was easier for me to actually write my data pipeline than write the spec to give it to the engineers to do. So I was like, yeah, this isn't so bad.

2:50So we realized very early on, like with these systems, you, I think increasingly as the tools, as our data becomes bigger, we're going to have new classes of product managers, including like data informed product management. I'm sure things like chat GPT are bringing that to the fore, but it's not just that. It's anything that adds a level of analytics to your dashboards and things like that. Like you want someone who has a business interest, but also is able to like run the SQL query to figure out what the hell went wrong with that dashboard. And so it's an interesting transition that I don't know whether I'm crazy for making it, but it's what the organization needs.

3:29So it ends up being a... I think it makes a lot of sense, right? Like if you're a data scientist today, you want to add that skill set so you can better interface with these tools that are able to extend and leverage your work. Correct. And it's been really like, so I taught myself Python mostly to sort of get a sense of how do we like build very reproducible processes to understand data? And then that's been a gateway drug into natural language processing, which obviously at Data People, we take career documents, whether it's resumes or job descriptions and sort of take and extract really important structured data so we can run analyses and understand the impact of things.

4:09So that's been my kind of gateway drug into things. But I sometimes read a little bit of PHP because that's what our lovely monolith is written in. And so it ends up being, once you know one of them, you start getting a sense of the others. You start seeing the patterns. Yeah, you start seeing patterns. I did not ever imagine myself as someone who could code or code competently, but it's kind of a, you find your ways. Because I think when you have a question that you want to answer and there aren't any other ways of addressing that question, you sort of are like, okay, let's have at it with all these different tools.

4:43And it's been a really great system. Is that a mindset that shifted for you over your time at Data People? It's now been what, seven years since you founded the company? It's been seven years. And finding the right tool and deciding on the architecture for those tools has been really important. I'm now starting to get a better sense of how systems are put together and kind of going end to end from the germ of an idea that you have to like building out, we're building out a second product right now. And that's been an interesting topic because the things that fuel the first product might not work with the second product.

5:17And so just splitting off infrastructure into very different systems just needs to happen. So that part of it is something that, honestly, a couple of years ago, if someone had told me I'd have to run things on AWS, I'd just be like, no, thank you. You guys keep it. I'm going to stay with my lovely statistics. But you get excited because you're like, this can change the nature of how we do recruiting. It helps teams do things in ways that are just so terribly manual. that like it just feels like a like I can imagine how engineers really enjoy that process because you set up a system and you kind of let it go and you see what people do with it.

5:59And that's amazing. So I'm very curious about what you're seeing in that data. Obviously, you now have this data engineering role where you're kind of bridging the gap between data science and engineering. Yeah. What are the trends that you've seen in your time at Data People and what are some of the big takeaways? One of the things that's been really, so I hate to be the scientist who said, oh, I told you guys so. But one of the things that's been very exciting in the last year is that, you know, we've been talking about things like title inflation for a long time. So an inflated title is when the job description doesn't match the title.

6:30So, you know, you have principal engineers who have job descriptions that are only asking for two years worth of experience. Principal engineers cover your ears. Well, I hope that they've all got more than two years worth of experience. But, you know, we've seen we've seen that trend. And we were annotating it, noticing it. And in the last year, it's kind of gone through crazy uptick. Like, you know, that's tied to the labor market. It's tied to companies trying to, you know, lure and attract candidates because it's really hard to hire in tech. So that's been something nice to see. I know I had a conference talk about it like five years ago or something where I was like, this is a thing.

7:06And people are like, what are you talking about? And now I'm so excited when I get to see. Like, look at the data. Yeah, it's in tweets and people are like re-liking those tweets and we're very excited. So it's actually, that part of it's been really fun. We're not the first to have thought of title inflation. Obviously, it's not our term, but it is something that we've been studying because we know that it has impacts. And now we have the data to show that, look, it doesn't matter what type of organization you are, whether you're a tech or a non-tech company, you've still got the same problem where at some point, like I like to think of that there is no free lunch, right?

7:37Like you can give someone an inflated title. If you don't give them the pay that goes with it or the responsibilities that go with it, they will eventually leave your company or they'll go somewhere where they will get that. That's been the lesson is being transparent and clear with candidates, with your employees is really important to having a well-functioning organization. And we just try and write software that helps people, you know, get there to help humans sort of communicate, which is wild. It's amazing how challenging that can be, right? Right. One of the things we think a lot about is like people don't love confrontation.

8:11Obviously, when you're collaborating with someone, having those hard conversations can be really difficult and it can be really easy to be like, here's a system and it's making us do this thing and we have to abide by it. It's much easier to blame a system than to like blame the fact that you two are going into a system, into a conversation with very different intentions. And so that alignment that happens in hiring processes really has to happen very early on. But it's the same thing when you're working as well. Like you need to be aligned with your manager about where you're going. You might have slightly different goals, but if you're over here and the manager's over there, you've got to somehow get together.

8:52So basically, yeah, it's not fun having systems that like force people to do things. It's much better to like facilitate that conversation and to come to a compromise. But yeah, we deal with those systems all the time. I'm curious to dive a bit more into the title inflation piece. Have you seen any impacts from the recent waves of tech layoffs? A bit of, I guess, deflation of that title. But it's still continuing at pace. We haven't yet seen the effects of that on the market. Too early. Yeah, it's too early. I think it's going to be really interesting. We saw. So in certain cases, we saw historically that deflated titles were really attractive to people of certain backgrounds, like especially if you don't have.

9:33For instance, an engineering degree, like being called a junior software engineer was like quite attractive to some folks that people like wouldn't think of. And it might actually be helpful for people's efforts to like. Try both. Especially if people are trying to ensure that they have like a representative candidate pools and things like that. We haven't yet seen it. I'm kind of curious to see how it's going to play out. Anecdotally, in our case, we're hiring a VP of Eng right now, which is very exciting. And the kinds of candidates that we're seeing in the pipeline, you know, like very senior SVPs at different organizations were not things we've seen for many, many years.

10:12And I'm sure we're not the only ones, but this is a weird market right now to be in. So we've had some similar experiences, I'll say, in the recruitment I've been involved in. So I'm very curious to see how that are kind of like anecdotal observations bear out in the data over that time. Correct. And I think like it's it's somewhat also restricted to the U.S. We've been tracking labor market developments. We do because obviously it's related to our product in different locations. And Europe has seen less of it. But Europe is also now seeing what the U.S. has seen for the last year, which was labor market tightness.

10:45So it's going to be interesting, especially companies, big companies that have like multiple places. How are they going to distribute work if you've got like more candidates in one place and less in others? But like the U.S. costs more. So it's kind of. Outside of title inflation, are you seeing any other major trends in the data? We're seeing quite a few. Our biggest trend that we're seeing right now is job descriptions are becoming incredibly long, which is fascinating from a natural language processing search process. because you know with algorithms like TF-IDF or other forms of normalizing, right?

11:23If you have a very long document that's got a lot of non-specific language versus a very short document that's quite tight, which one's going to do well on search? Which is a lot of companies can talk about their search things, but basically it comes down to some level of normalization of a link. So we're seeing that link and a lot of companies talking about their values and their job descriptions, like coming back to, like there's a big push about culture and values that's happening. I don't know how that's going to pan out. We have these discussions a lot and some of the stuff is coming out because of pay transparency.

11:54So in Washington state, the regulations require you to disclose exactly how many days of leave you'll have or any non-comp benefits that are like meaningful. And so that's changing the nature of job descriptions. What I'm super curious about figuring out is what's happening to candidate pools and things like that. people are talking a lot about like, you know, you'll have multiple locations and things like that. But that's the big trend that I'm sort of curious about. It's not the inclusion of salary, but how does, like a lot of actually job seekers are like, I don't believe the salary that's in the job description, which is sort of interesting.

12:33That's intriguing. Or is the salary above or below like the median? So can salary act as an attractant? Like, do you trust that as a signal because there's like, There's a new era of transparency, I think, that's happening in job descriptions. It is fascinating. I'll say I'm a Washington State local and I work remote, but I see it when I'm getting the LinkedIn sponsored posts of like, oh, apply to this job. It's been really intriguing to see that trend change or suddenly now there's a lot more upfront information. I'm also seeing, I'll say, like very broad ranges. Oh, yeah. 90 to 900 ,000. And you're like, how?

13:07So I've spoken to HR leaders at organizations where they've got like their tech companies So they have this sort of thing and they're like, they're very worried about being litigated. And so they're like, if we give them the possible possibility of range based on historics, including a lot of other information, then we won't get sued and they don't want to get sued. So they're very afraid of that. It's understandable. Washington State Attorney General Pop Ferguson is very willing to enforce regulations. It is. But on the other side of things, you're like, you're you're not in the spirit of law.

13:42And I think in Oregon and in New York, they've been sort of, I think you see the ranges are much narrower because they're like, it's in your best estimate. So if you hire someone at a different level, a slightly higher level, you won't get penalized because it was your best effort. And you're like, you know, it was an L6 versus an L7. But in Washington state, it's not, they haven't been explicit and they haven't gone through that whole cycle. Probably some regulatory clarity they need to bring to that. They do. And I think you'll see that pan out in the next year and hopefully that will add clarity.

14:13But it's amazing because one of my teammates, she's she used to be a teacher for a long time. She's she's a linguist on our team. And she's like this whole thing, like we have complete clarity in every other profession. Like she's like, you know, her parents sort of work in warehouses and production roles and like salary transparency has been part of that industry for long as well. But she knew what her colleagues were making at her school as well. And she's like, you knew what you needed to do to get to that next stage. So she was like, this is wild that there's a lot of, there's so much angst because it's like, it's fine.

14:45It's a, it's a big change, but it hasn't fundamentally changed other industries that much. Like the things that people were afraid of haven't yet happened. So, so yeah, that's been an interesting intersection where we've seen like things change in the last year that like I never thought like I was just amazed when it got signed in California because I was like well that's gonna change things for a lot of big tech companies all of a sudden and they also have to report to the government what their gender statistics are in terms of say salary and like levels and things like that which is a which is the other part of that requirement that I don't think anyone's yet really talking about too much yeah but people have been doing in Europe for years.

15:26So it'll happen. So it'll happen. This is fascinating. Sorry. I'm like, I get super geeky and excited about this. No, I want to know more trends. I'm like, what else is happening? What else is happening? So we have titles. We have... For folks listening, I've gone entirely off scripts. We had an outline. We're talking about these things. I'm like, no, this is so fascinating. What else have we been thinking about recently? I'm just trying to think of stuff that I've talked to our customers about because we have these conversations with customers to talk about like what's happening in this stack how are they changing because these are recruiting teams at many companies they've kind of gone through significant layoffs because if they're not hiring economic times they're not going to be dealing with those with outside of it so i think the other thing that's really interesting in in the our field that hasn't yet happened everywhere else and it's it's sort of like i think of recruiting as where marketing was like 15 years ago where it's wild to me that we spend the u.s i think spends 17 billion dollars on executive search every year like that's the size of that market where it's like a headhunter or reaching out to someone and it's not just like senior level roles where that's happening with see that's happened in junior level software engineer roles.

16:49The cost of like getting interns is sometimes more than the amount that you'll pay the interns in terms of not acquisition, but like, you know, you'll advertise in places. And I think what I think is a really exciting trend is unfortunately caused by the economic situation. We need to like people need to be able to talk about efficiencies in the process. And that's where their focus is. And so that's causing recruiting to sort of be like, what are the costs of all of our channels and which ones are actually doing well for us? It's becoming a lot more data-driven than it's been in the past. And so what I love about recruiters is they're like, our hiring managers, our engineers, they want us only to advertise on Stack Overflow or GitHub jobs, but we don't get any applicants from there.

17:33They've been like, we want to know about niche job boards because we want to go back to the people and say, this is the evidence, this is the cost. We could pay a candidate$100 per candidate and it'd be cheaper than advertising on those platforms. And that's wild to me that you could pay to get so little and have no visibility into that process. You start applying more. What am I doing? Yeah, I know. Like, I'm like, if you just put our Craigslist ads for applications, it might actually be cheaper. So, I mean, I think there's a reckoning because this whole field has kind of got to go through that sense of like, what's the cost of, not a cost of a candidate, but like, how do we get to more candidates more efficiently?

18:10Candidate acquisition costs. We can just say CAC again, We can say CAC again and it's like, what's their lifetime value? Like we haven't kind of gotten to that level of sophistication. But I think it's an intuitive step. I think one thing that I am always amazed by though, even though marketing has had this idea for a long time, is that their stacks are as bloated as everyone else's. Like it's like I hear of so many point solutions that do this thing to this tiny little step, but they don't take you the full way around. Like we, our marketing stack is as bloated as anyone else's. And that's an area where I think recruiters can be a little bit or recruiting teams can be a little bit more efficient.

18:51The other thing that's actually really interesting, that's, this is the other trend that I think is that impacts candidates less, but it's actually fascinating. So business systems, if anyone's ever dealt with an ERP, an SAP, a workday or whatever else, I would call them user hostile systems, like really frankly. and they are horrible to navigate and they don't invite collaboration, right? Like you have your workday person to make workday a little bit more pleasant to deal with. And it's just amazing to me that you hire someone in your company to like create this like layer in the system that is just, I don't know if you've had experiences with these things.

19:30I mean, we hire Salesforce engineers. Right, exactly. Salesforce engineers, like HubSpot administrators. Like it's like this whole, like for me, the idea that those software softwares need someone in-house because this tells you a little bit about like I come from academia we don't normally like everything is self-service there so the fact that you have in-house people that are experts in this one thing is is amazing to me and that we call it still a service because I'm like it's kind of a bit of both we have to hire someone to manage the service yeah and like maintain it it's scale sure sure but like it's a tense thing.

20:02So these are user hostile systems and you generally don't want everyone in them, right? You want to train a group of people who can deal with this user hostile system so that your hiring managers don't have to touch it. They can sort of have the experience for the candidate. And what we're seeing recently is that this is a collaborative process. And the more that we make these systems collaborative and able to be interacted with together, the better that like deals with that alignment issue that you don't have someone putting some set of things into the system where the other person is like, you get to an office stage of the candidate and the hiring manager is like, who the hell is this person?

20:39What are you doing? Like, based on what? Like, bringing them, having systems of collaboration. I think Google Docs did an amazing thing for collaboration. No question. Like that whole industry has done an amazing thing for collaboration. We have that in any of those, like, can you imagine doing that in Workday or Salesforce or any of these systems, that's where I think the next generation of like business software has to come, whether that is within Salesforce as like an experience layer or whether it's like, hey, I'm going to burn Salesforce down and build you this new thing that allows you to collaborate and interact in a way that's like part of the system will be different.

21:17So I think we're seeing that trend in terms of for us, we integrate with a lot of application tracking systems. So things like Workday, Greenhouse, Lever, for those of you who are in startup land. And like they're getting a little bit, they're better than what the previous generation were. But I think there's another generation that will have a much more consumer focus to that business software. And I think bringing people into that process is going to make those processes better rather than being like, here, recruiter, you deal with your little sandpit and I'll deal with mine. And then we'll collaborate over Outlook or Google Meets or whatever, like as in whatever your messaging system is.

21:55How are you applying machine learning to this massive data set you have? Because I'm super curious to see what insights that's generating and how it's fueling your product innovation. So we use language models in a couple of different ways. One is to sort of extract entities. We've been training that, for instance, to pull out salaries from job descriptions. Part of that is because we want to tell companies, our customers, if they're not compliant with that legislation, then that's been actually really fascinating. Oh, that's an awesome tool to have. Yeah. Okay. So, but we've been doing that for like, ever since the salary legislation came out, we were like, yep, we should be capturing this and trying to surface this up.

22:30And that's been really an interesting experience. The other area where we use machine learning is, you know, we extract a lot of features. Our features are not unsupervised. So we do have some supervision because we think that things like title inflation are really subtle things that like make sense to a human that are hard to do that. But what we then do is throw it all into a model and sort of say, OK, what are the things that are like most correlating with success in this process? What are trends that we're seeing, which I think has been a fascinating side of it? Like so a really good example is we published a tech hiring report last year, at the start of last year, talking about how skills and qualifications and even titles in different tech jobs have changed over the last couple of years and like how we're seeing specializations develop in certain types of roles.

23:19So companies are less likely to talk about, certainly more startups are less likely to talk about software engineer. They'll be talking about a front-end engineer or a Vue.js engineer versus like a back-end engineer and being kind of specific about the subtypes that that's definitely happening. So like we're seeing, I guess, I wouldn't call it maturation, but it does feel like it's like, I think in biology terms. So like you start with an embryo, you end up with a very, but your hand, right, which is like full of very differentiated cells. And we're getting to that level in some fields within tech where we've got that level of specificity and others are still nascent.

23:55They're still growing. So, you know, more companies are adopting things like analytics engineers than they used to be. It was a very startup thing to have. But I think now people are less likely to be BI engineers, more likely to be analytics engineers at different jobs, actually. But it's been it's interesting to see that rise. What was the sort of subset that we saw a couple of years ago? Very handful of companies had where it's now like, you don't have an analytics engineer. What are you doing? Like, of course, you need this person. Are you going to do an update to that report this year? Yeah, I was actually thinking of doing it.

24:28It was such it was actually a lot of fun to generate. And it was funny because we did do an update at one point, a small one for like a Business Insider article. And what was amazing was whenever you haven't touched infrastructure in ages, you think you're like, it could take five hours. It could take five weeks. I don't know which of these it will be. It was five hours. We have a great engineer and it ran really well and it didn't break. And I was just like, I don't know how well we engineered this, but he did a phenomenal job. So, yeah, we will because we're like, yeah, we've got a whole lot more data to kind of look at.