Breaking Into Data Careers: Practical Paths for Newcomers and Career Switchers with Avery Smith of Data Career Jumpstart

In this episode, Avery Smith, Founder at Data Career Jumpstart, joins us to explore how newcomers and career switchers can break into the data world. He shares why most people start with the wrong tools, why data analyst roles offer the easiest entry point, and how focusing on skills, portfolio and network creates real momentum. Avery also breaks down why internal pivots often outperform external job hunts and how AI is reshaping, but not replacing, the work of analysts and data scientists.

Key Takeaways:

‍

00:00 Introduction.

02:37 Beginners shouldn’t start with Python because it adds unnecessary complexity early on.

06:07 Data work should prioritize real business impact over flashy tools.

10:11 In tight markets, companies prefer analytics because it delivers quicker, more reliable wins.

15:13 Skills matter, but your portfolio and network are what actually create opportunities.

21:19 The real value isn’t the code — it’s the insights it produces and how clearly you show them.

26:14 Lead with small asks; advice opens more doors than asking for a job.

33:03 AI has assisted data work, enhancing workflows rather than replacing roles.

37:15 Good data still needs clear explanation to drive real decisions.

‍

Avery: [00:00:00] one like easy way to. To land a role that maybe you're not even qualified for is to pivot internally. That's the only reason I went from chemical lab technician to data scientists because it's, 'cause I didn't switch companies.

So my company was like, oh, Avery's already a really hard worker. he's already proven his worth. So I personally rather always try to do an internal pivot because you don't even have to apply for the job.

The job kind of applies for you.

Anthony: Welcome back to Data Masters. Today's episode is the ultimate guide on how to break into the data world. Whether you're a fresh graduate with a blank resume [00:01:00] a mid-career professional looking to pivot after years in a totally different industry. My guess is Avery Smith, a top 20 data science influencer and founder of Data Career Jumpstart. He's built a massive following, helping everyone from high school math teachers to music therapists, land six figure data jobs. And we're gonna cover the specific playbooks for both stages of life, both new grads, where he argues you should ignore GitHub and focus on building a portfolio that tells a real story. And we'll also talk about career switches, how to weaponize your past life experience to beat out computer science majors. and for everybody, I think, Avery has a controversial take about why starting with Python is a big mistake. Avery, welcome to the show.

‍

Avery: Thanks. I'm glad to be here. Thanks for having me on.[00:02:00] 

‍

Anthony: Cool. So let's start with that controversial take. you argue that. Most people trying to break into the data world are starting with the wrong tools. And, we've, talked a lot on this show about Python and about why, you know, the world of data is so anchored on this idea of writing code and, and specifically very much writing Python. but you don't agree with that. And in particular you have a different set of. Tools or a different data stack you suggest people start with. So, share your, your take.

‍

Avery: I should start by saying that I love Python. It's my favorite tool to analyze data and I think it's amazing. But like you said, I don't think if you're just starting out, You should really be learning it. And the reason why is there's just so much more low hanging fruit than Python. when you're learning data, that's a lot of work.

You have to like learn different types of charts and how to clean and how to [00:03:00] analyze and how to do statistical models and how your data. And that's a lot in of itself. And then on top of that, if you're going to learn Python, you're also learning about programming, which in itself is like a whole language and a whole, a whole thing separates to learning the data concepts.

And so to me it feels like, man, that's a lot to learn how to program. And all these data concepts at the same time. And so my point of view is like, okay, what if we just simplify it? We just focus more on the data concepts first. and we do easier tools like Tableau or Excel, which you're already familiar with, or even sql, like sql, Is hard as well. It's a programming language, but it's kind of easier than Python to be honest. There's just, a lot more structure to it for the majority of the time, and you can honestly do less with SQL than you can with Python. So I'm a big fan of Python, but it just feels a lot, it feels very overwhelming when you're just getting started.

So I'm like, okay, let's just start with something a little bit easier and work our way into Python down the road.

‍

Anthony: Right. I think there's also the sense that [00:04:00] in when you focus and start with Python, are in a way jumping over the real work of being a data analyst, which is shocker analyzing data. So you get wound around the axle of trying to. You know, get your Python code to be syntactically correct and lose sight of the question that you're trying to ask and answer with the data.

‍

Avery: A hundred percent. And, I actually, I have a job board that I run, and so I have access to all of the, the job descriptions. And so I've gone through, what. You know, 10,000 job, descriptions. Now, these are for what I would consider data analyst jobs. These aren't like data scientist jobs. So these jobs are like, I consider financial analysts, pricing analysts, business analysts, healthcare analysts, business intelligence engineer.

I consider all those kind of derivatives of the job family, of, of data analysts. And when you look at that, Python and the number of times it's actually mentioned in the job descriptions is behind. sql, [00:05:00] Excel, Tableau, power bi, then Python, so it's fifth and it's only required like in 16% of the jobs.

So it's like, it's just, there's so many things you can be analyzing without Python and there's so many jobs you can be qualified for without Python. That the learning curve, which is quite steep for, for Python. Just, I just don't know upfront if the investment in learning Python is worth it. I do think in the long term.

It's worth it. Like I think you should learn Python eventually, but when you're just getting started and you're struggling to to land a job, I just don't know if it's the best investment.

‍

Anthony: Yeah. And there's not a huge number of things you. Can't do in a tool like, Excel or, or or BI tool that, you know, clearly there are some things you can do in Python, but even, I mean, taking a simple example, you can run a regression in Excel. the real questions. are you running the regression? Like, what question are you to ans answer that, that, that that technology is [00:06:00] the appropriate solution to the problem.

And if you've skipped over that, then you're just, in my example, running regressions for regression's sake.

‍

Avery: We don't wanna be doing data for data's sake at all. Like, and, and also we don't wanna be. In our analysis, we don't necessarily wanna choose the sexier way to look more sexy, right? Like, like you said, like we just, we need to care about the business results, the organization results. Whether that's saving time, saving money, saving lives, like that's really what we should be in data for, not for doing Python.

'cause it's fun, although it is fun and I think, I think it's worth learning in the long term. But yeah, a hundred percent. We don't wanna just be doing data for funsies.

‍

Anthony: So we've also, I think so far in the conversation, been a little loose in our language. We've both used the term data analyst and data scientist, and we've, at least I have been mixing the two up a little bit. I think there's a, a sense that, being a data scientist sounds cooler, who wouldn't want to be a scientist?

and. Presumably, or you could probably tell me [00:07:00] with, with the job board data, whether it actually pays well. but you spend a lot of energy encouraging people, especially initially to go after data analyst jobs. And so maybe start with what's the distinction between these two roles? and then following on why your view is that we should be, people should be going after these, analyst roles to start.

‍

Avery: it's interesting the data world, all the data job titles are a little bit fuzzy, so there's gonna be some overlap. in fact, I had, I think his title, he. It was one of the bigger car dealerships, like this car brand you've definitely heard of. And I think he owned a couple dealerships, on the east side of the United States.

And one day he reached out to me on LinkedIn and he is like, Hey, I'm, I'm posting this job description. Will you take a look at it and let me know what you think. Maybe you know, someone who's a good fit. And, he had titled it data scientists, but like all the descriptions were like data analyst. and so there, there's definitely people in jobs out there who, who conflate the two.

but for me, the definition. [00:08:00] Basically most of the time what direction you're facing. I think that's the easiest way to look at it. Data analyst is going to be doing more of like the descriptive and the diagnostic analytics, so that's like what happened in the past and maybe why, and a data scientist is going to be looking more in the future.

Predicting the future predictive analytics. So what's going to happen? And then prescriptive analytics, how can we make it happen? so yeah, it's a lot easier to tell what happened in the past than it is in the future so that data scientists often do get paid more because that is a more difficult skillset to, to do.

so that's my definition is just kinda like, what way are you facing? Are you just kinda looking at what happened in the past or are you kind of looking more at what's gonna happen in the future? Or are you trying to create models to. Get an ultimate outcome that you actually want. So a good example of what a data scientist would do is like, you know, what order do we show your TikTok videos?

What, when you open up Netflix, what are we actually showing to you? That's like trying to get users to stay on the platform, and that's a data scientist problem versus [00:09:00] a data analyst problem, in my opinion.

‍

Anthony: Yeah. I love that distinction. 'cause I think it's. So easy for people to understand. And I also, acknowledge and appreciate this distinction of your drawing between the job poster, if I can say it that way. I'm the job, receiver. 'cause I do think there's this sense that, people who are posting jobs, like, why wouldn't I want a scientist over an analyst like that sounds better. but if the. Problem I'm trying to solve is a descriptive problem. If I'm trying to understand my business better, I could almost certainly get away with somebody with a set of analyst skills. Is that the right way to think about it?

‍

Avery: hundred percent. And, on my podcast, I recently just made an episode about why wouldn't maybe necessarily try to go from scratch to a data scientist in the next year. Or so, and one of the reasons is I think we're in a really interesting, economy right now where like s and p five hundred's still going up, but like.

We're feeling kind of tight and it's like what's happening? And I think a lot of [00:10:00] companies, that maybe don't have, like, let's say that don't use machine learning as like a core value proposition. So like Netflix, like that's really important to them. Instagram, it's really important to them, like the algorithm.

Anything that has like an algorithm, they're gonna need machine learning and data scientists are gonna be doing the machine learning. Let's just say ignore, like those more techy companies. I think a lot of companies are gonna feel. Tight and for me, you can get quicker wins with data analytics 'cause it's just easier to do.

Like looking at what happened in the past. A lot of companies aren't doing a good enough job at that right now, and it's a quick win for them. Data science where you're trying to break the future, trying to be something a little bit more difficult. Like a lot of data science projects fail, and I think when companies are like, feeling kind of stressed, you don't want failed projects.

So, I kind of, I'm not sure about this, this is just my prediction that like data scientists roles might stall a little bit next year as we kind of feel in this tight environment while like data analyst roles, it's, it's a little, a little safer bet. I guess you can get, a little bit more of a, a proven [00:11:00] ROI.

The ROI might not be as big, but it might be more stable.

‍

Anthony: And so that, and that's from the perspective of the employer, but thinking from the perspective of a new hire, or sorry, a new grad, rather, somebody who's just entering the workforce or to the earlier, or to where I started us. somebody who's mid-career making a switch, I guess you would say. Positioning as in that analyst vein. There are more jobs. They're easier to get, they're easier to do, and it's, you're gonna be more successful in that first role.

‍

Avery: that's exactly what I believe, and I learned this the hard way actually. because truthfully, I kind of went from being a chemical lab technician to a data scientist. I kind of jumped straight into the data science position and. And, when I was like, man, this is awesome. I wanna start teaching other people how to do it.

I started like a, a bootcamp and I, I ran it for a year. And, I had mixed results. I had some students who had landed jobs and some who had it. And when I did the analysis and looked at like, okay, you know, [00:12:00] who actually landed data jobs? Was it the smartest people? Was it the people who were best at SQL were, did they have the best Python project?

The answer was no. It was the people who were aiming for the data analyst roles. And so going from zero to data scientist proved really difficult. And so I ended up changing my whole philosophy of like, okay. I'm not gonna teach people to go from zero to data scientists. I'm gonna teach 'em to go from zero to like one of these data analyst roles.

And then from there it's like the hub of the data world, almost the tech world, like where you can become a senior data analyst or a data scientist or a data engineer down the road, but like, let's just get your foot in the door the easiest way possible. And once you have a data job, life is easier.

‍

Anthony: I love that. And actually there's a, that's a really nice way of thinking about it, which is if you want to be a data scientist, get your foot in the door as an analyst and then prove your. Prove your metal right, versus getting to a tough situation where you've taken that data scientist's role and maybe you're not ready, you don't have the skills, and certainly you haven't proven yourself within that organization, and that's a recipe [00:13:00] for a, you know, a tough conversation.

‍

Avery: Yeah, that's not a place I'd, I'd, I'd rather, I'd rather work up. and, one one like easy way to. To land a role that maybe you're not even qualified for is to pivot internally. That's the only reason I went from chemical lab technician to data scientists because it's, 'cause I didn't switch companies.

So my company was like, oh, Avery's already a really hard worker. I was super underpaid at the time. So they're like, we're great Val, he's great value. like yeah, that's, this is a change that we're okay making because he's already proven his worth. So I personally rather always try to do an internal pivot because you don't even have to apply for the job.

The job kind of applies for you.

‍

Anthony: Well, and very much to your earlier commentary about where perhaps the economy's going. these sound like a, a super smart, super grounded, strategy. So you talked a little bit about the coaching business. You've been developing. you also have a very specific methodology, which I love this idea of the SPN method, skills portfolio and network. and you also make this point, [00:14:00] which I appreciate, and is that people. Overemphasize or over, spend energy and time on the skills portion of skills portfolio and network. And you argue the other two are actually not just equally important, but in fact, more important. I might add that I would, this potentially could be a function of the fact that we get a lot of kind of. Overachieving people who've spent a lot of time in academics as students where skills and skill assessment is really the, the primary, mechanism of, of getting feedback. And so no wonder they spend so much energy sort of building skills, but enough about, tell me, start, go backwards. Tell us a little bit about the, the methodology and, and why you think, the p and n are more important than the s.

‍

Avery: Yeah, I wish we lived in a world where it was just like the most skilled person gets the job, but it's just not how reality works. You work on it, right? That is kinda how school works. I never thought of it that way, [00:15:00] but in the real world, it's not the, the best product that is the most popular. It's not necessarily the best singer.

Who has the most streams, right? There's other factors to it. And the same as landing. Any job, especially data jobs. So your skills are, are important. You need to have the skills, but the way that you're presenting your skills and the opportunities you have are coming from the, the other two parts of the methods.

So there's the S skills important, the portfolio. P and the network n are the other two, parts of the equation. The portfolio is what actually is like the evidence that you can do what your resume says that you can do, because it's like, if you're like, yeah, I can, you know, I can code in Python, a recruiter or a hiring manager, especially when you're a new grad or you're a career switcher is gonna be like.

Okay, well where's the evidence? Right? Like, show me, show me the money, show me where you can actually prove it, and you're like, I took a class or something. Right? A portfolio is more tangible. It's like undeniable. Hey, look, your job description asked me, you know, asked for, can someone. Look at marketing data and be [00:16:00] able to choose like the most optimal, channel to increase ad spend on.

I have a project where I literally have done this for a different company, and the recruiter and hiring manager have to be like, whoa, you're right like this. I can't deny what you're showing me. So it's the, it's the evidence. And then even if you have the evidence, even if you have the skills, if you don't have.

The network or you don't have the, the job application and the career skills to get yourself in front of the right people. It's not gonna, do you any get, like for example, let's say I, I, I'm in my office right here, right? Let's say I cured cancer. Absolutely amazing. If I don't tell people about it. Like, nothing's gonna change in the world.

I'm not going to make an impact. I'm not going to be able to capitalize off of my invention like you need to be in front of the right people. You need to have the right opportunities come before you. And if you, and if you don't actually make an effort to do that, it's not gonna happen.

‍

Anthony: Yeah, so I sometimes think about the skills as if it's a poker game. It's the ante, like it's enough to get in the game. To your point, if you don't have them, [00:17:00] you're not playing that hand. You're sitting out. So it's, it's not. A, an excuse for. Ignoring the skills, but it's, it's also important to recognize to very much to your point that the most skilled person doesn't get the job.

But let's dig into this portfolio question a little bit. what is the best way for people to show off their portfolio? Like, do I, create a bunch of GitHub, repositories and send people a link to, to GitHub? Or how do I get people to know that I've done this, done this thing?

‍

Avery: Let's, let's split it into the two categories, the new grads and the career switchers. if you're like a current student or let's say you're a new graduate, what I did when I was in college is I literally had a binder with pictures of like, Hey, I made this graph. Hey, this is like the, the end results of this machine learning algorithm I made and I would physically show people in person interviews.

I think that's really impactful. 'cause once again, it's more tangible. It's like, Hey, look, it. You literally can't ignore this. It's right here in front of you. so I think that's a [00:18:00] fun way, but I think a lot of the world and the way we're moving is obviously more digital. So I think having some sort of online, repository, of your projects that's, that's tangible is good.

There's lots of different options. You, you can do. GitHub is probably historically the most famous, but I have kind of a hot take where I actually think when you're first starting out, let's say you're not actually coding that much. 'cause, 'cause GitHub is built. For collaboration between coders and version control.

It's not built to be a portfolio. You can hack it to be a portfolio, but it's not actually built that way. And in fact, I think GitHub admits this because they built a new product called GitHub pages that is more of like a showcase, type tool inside of the GitHub platform. So I'm personally not a huge fan of GitHub.

I think it's a little bit clunky to the portfolio, but like doing something like GitHub pages. Or using Squarespace or Wix or card is what I recommend a lot of the time. just these simple website builders, drag and drop website builders, I think those can be great. but you could even do it with like LinkedIn.

I think LinkedIn. you [00:19:00] can kind of do some workarounds in LinkedIn. Where you can have, they have a project section you could utilize or my personal favorite is we, we create what's called LinkedIn articles and then we basically pin them in our featured section on our LinkedIn profile. So if you open up one of my students' LinkedIn profile, you'll see their profile picture, their headline, their about section.

But the next thing you see is like. Seven pinned projects, at the top, basically like this is my SQL project, this is my Excel project. So once again, trying to get the evidence in front of people as quickly as possible.

[00:20:00] 

‍

Anthony: also make an important distinction there, with your example of the. printed graphical output. And just to contrast that with GitHub, I think another reason GitHub doesn't work is you're asking the reader to do the work. Like, here's a bunch of code I wrote. should assume that the result of all this code is this great insight. Almost by definition the code isn't run, it's, it's the code. in contrast, what you're showing with the chart in your printed example is, the outcome of this analysis is this set of insights that I've. Clearly summarized for you, hopefully visually. which by the way is what your customer wanted.

They did not want your GitHub code. They want, or their code. Let's just not pick on GitHub. They don't want your code. They want the output of the code. And when you walk into an [00:21:00] interview and you show the code, you're effectively telling the inter the interviewer, you need to do some work to believe me.

‍

Avery: A hundred percent. It's really hard on, on gab. You have to like basically make a really good read me file and like screenshot a bunch of stuff, and upload the output onto gab, for example. But yeah, a hundred percent. Once again, going back to we're not writing SQL code to write SQL code, we're writing SQL code for the, for the business decisions we're making that the code outputs.

It's really the value in any sort of. Python code or SQL code is not, not even the code, not even the results from the code. It's the insights derived from the results. So if you can, if you can try to show that you can do that, all three of those things, I think you're gonna put yourself a leg up against the competition.

I.

‍

Anthony: And, and then, maybe to, to be this one to death a little bit, the visual dimension. I think there's also this tendency, and this is admittedly a bit of a psychosis of data analysts and data people, is they want to give answers as data. [00:22:00] And you know, my experience of most executives is. A wall of numbers is overwhelming and boring and a bar chart, you know, or whatever. hopefully not a pie chart, but a bar chart is like a really good way of saying like, here's an insight.

‍

Avery: Yeah, it's really powerful. The visualization's a great thing to, to make your projects really stand out.

‍

Anthony: So, let's shift a little bit to this idea of the network. so in a way that's maybe, a bit more of the surprise to me. think one, angst that frankly, you know, whether you're a new grad or a career. transition. I hear this most acutely from new, new grads. 'cause they feel like they don't have a professional network.

I mean, almost by definition they've been in school. And then the career switcher almost certainly would say, have a network, but it's the wrong network. I'm trying to get into data. My network is in chemicals In your example [00:23:00] the industry is. I'm, I'm in. And that doesn't help. And so I'm starting from this sense of like the lonely person on an island in the middle of the, the ocean. how do you think about the network?

‍

Avery: I'm gonna try a new concept, on this podcast, so hopefully I haven't, I haven't thought this all. The way out. Hopefully it goes well. But there's really, there's only really four different ways to network and there's two variables on, on each way. So think of it as like a matrix. You can network one-to-one or one to many, and you can network privately or you can network publicly.

and there's, so there's kind of four different combinations here. the, the easiest. And well, actually the easiest really depends on your personality. It all, it all depends on you and what you're comfortable doing. And I try to teach all four of these methods in, in my bootcamp and then students can select which one matches to them.

It's like one really easy way, in my opinion to network one to many. Publicly is to post on LinkedIn to just document your journey and then all of a sudden people will start to notice you. I started just to talk about data [00:24:00] when I worked at, ExxonMobil. So I was already a data scientist. I was already in the industry.

I was just talking about what I was working on and what I was studying on my own. and that changed my life because one day a private equity, I don't even like VP of a private equity company reached out to me and was like, Hey, would you ever wanna do consulting for us? And I was like. yeah, sure.

Like that's great. And so that's kinda what started my freelancing business, which allowed me to quit my corporate job and kinda start my own business. so that, that works really well. But if you're like, oh, I'm terrified of posting on LinkedIn, that sounds really not fun, then maybe like doing like more one-to-one would be more your revenue.

So like, for example, one really easy thing to do is just like, look to the neighbor that lives to the left of you. Look to the neighbor lives to the right of you, and just write down what they do and where they work, and answer one question. Does their company hire data people? If so, then reach out to them and be like, Hey, I like, I know you work for this company.

Do you like it? That's like such an easy question just to start the conversation. and then eventually you can be like, [00:25:00] do you know any data people at your company? And the answer could be yes. And then you can be like, oh, well can you introduce me to them? Can I maybe talk to them about like what it's like The answer may be no.

And you can just say, well, do you know, like a recruiter who might know someone that does data at your company? and those two avenues I think are, are two, just depending on your personality type. Two really easy ways, no matter where you're at in life to start, do some sort of networking.

‍

Anthony: Yeah, and I think there's also this sense that, networking needs to result in a job with the person that you networked with on the first conversation you have with them. You know that very day, I think disusing yourself of that expectation, ratchets the pressure down. So I love this idea of reaching out to someone and saying, do you like your job? for one thing, there's probably an insight there, which is, you know, maybe their company they work for is horrible and you wouldn't wanna work there. So just immediately learn something. But more likely they like their job and they're excited to talk to you about it and want to share that insight with you. What you didn't do is say. know [00:26:00] you work at Company X. I'm trying to, you know, get into data science. So data, be a data analyst, here's a job posting. Can you get me this job? You know, like that's a heavy, tall ask, whereas do you like your job is a low ask?

‍

Avery: Starting with a low ask is, is really important, because it brings the pressure off I think of everyone. I do think though, like if there's an open job at the company and you're actually tight with this person, like they, you are actually friends, I think it's worth reaching out there. But yeah, if it's like someone who's kind of random there's a really good, maxim in like career searching and career advice, and it's if you ask for a job, you'll get advice in return.

If you ask for advice, you'll get a job. And so if you can really start and be like, Hey, what would you do if you're in my situation, can really open the, the door down the road for those types of jobs.

‍

Anthony: A hundred percent. the other thing I always sort of suggest, so you, I, I love this distinction between, uh, one to. Knee versus one-to-one, and public versus private. The other piece that I often [00:27:00] suggest to people is. and every conversation with a question about who else you could talk to. So if you think about it as like a network effect, you talk to one person, they introduce you to two, each of those people introduced you to two. Soon you'll be talking to every person on planet Earth. Just mathematically we can work it out like it. This, these doubling factors are, are very powerful. Even if that person introduces you to one other person, like that's still powerful. So, you know, think about it as. Building this tree of relationships where in a way almost you're not expecting any one of them to turn into a job.

What you're hoping for them to do is to turn into another conversation.

‍

Avery: Yeah, that can be really powerful. I had a, I interview. Had a guy on my podcast maybe two years ago, and I can't remember exactly why he cold messaged, this person at this company if he was just doing informational in interviews. Just like, Hey, like, do you like your job type thing? Or if there was already a job.

I don't think the job was up. I think he basically [00:28:00] reached out to this person and was like, I wanna learn more about your job. I'm interested in your role. Would you be willing to spend 10 minutes with me? And then he asked that question at the end, Hey, who else should I meet? and it was, oh, you should meet my coworker.

And, eventually when a job did open up at this company, he had already talked to like five of the data analysts on the team. And so when the job opened up, he applied, they saw his name and he was like, all, we're just gonna hire this guy. And it was easy from there.

‍

Anthony: Yeah, exactly. And it, it, it, I think these things are very easy. To when you look back on them, like it's very easy to draw the path from where you were to where you are today looking backwards, but it's effectively impossible to, to draw it the other way, the serendipity that you know, that, that these kinds of relationships, you know, that, that the, the hops that you will, that you take through the network, you know, you'll never figure it out.

Expo. You have to look at it backwards.

‍

Avery: A hundred percent.

‍

Anthony: So you've been doing this for a while. sometimes I think, you know, we've been talking in a way, quite theoretically, maybe share a [00:29:00] story. I think you have such great stories about, you know, people, you know, maybe help people who are listening feel the art of the possible.

Like, yeah, like, I'm a, I'm doing this and I want to get into the data world. I think it's interesting, it's the future, et cetera. like, you know, are there, I, I'm sure there's some great stories of, of people who've made a transition that you've helped.

‍

Avery: I, we, we, we try to help as many people as, as we can. We've had some pretty remarkable changes and I try to share 'em on my podcast data career podcast as well. but I'll, I'll share some snippets. Some fun stories with you quickly. So, my favorite people to work with are actually like high school math teachers.

'cause they know math already. So the math and data is not intimidating to them. They're great communicators. And being a teacher, and I can say this 'cause my mom's a teacher, being a teacher kind of sucks. so like. The bar to beat, a teaching career. It's, I mean, they're amazing. We need teachers in the us but like, you just have to dedicate more than 40 hours a week and you have to be in person.

so I've had a high school math teacher go from making, I don't know, let's say [00:30:00] $75,000 probably working, you know, 40 to 50 hours a week to. About 120,000. so not quite double, but pretty close. after my program, and once again, she was very skilled. She went through my program. She built the portfolio.

but what's the third part? She needed a network, so she actually somehow knew, or her cousin knew a recruiter that worked for a really large bank. And so she ended up connecting with this recruiter and landing this senior data analyst job at this large bank institution. so that was awesome because like, just imagine like.

All of a sudden, oh, and she works, she works hybrid in the office, I think once a month. So I mean, imagine going from basically, let's just call it doubling, even though it's not quite, but, but almost doubling your salary, like extra $50,000 a year. All of a sudden you don't have to commute to work anymore and you're doing something you really enjoy and you never get in trouble with parents or, or principals or something like that.

Like, that's like a huge win. and it's totally in reach for not only teachers, but a lot of different [00:31:00] professions. Who don't really like their job right now?

‍

Anthony: I love that. And you know, exactly like, yeah. And, and I also particularly love the, what was it, the cousin of the, you know, person at the company like, and it's a perfect. Example of that, you can't plot these things out going forward. You can only tell that story looking back on it, right? Because you can only be like, oh, this is, of course it worked out that way.

And so, you know, but if you'd asked her, you know, before she engaged, like, yeah, in your family will familial network it works, you know, is the right person for you? She'd be like, I have no idea. Right. But you gotta kind of go through the process. So I wanna, I wanna ask you a tough question. it is sort of the, the elephant in the, the room we haven't talked about, which is, and which I feel like has to come up on every pod.

It's like a, a podcasting rule that we

‍

Avery: Yeah.

‍

Anthony: up ai.

‍

Avery: Yep.

‍

Anthony: but, but I think there's some, particular relevance in this context. Because context, think that there's a [00:32:00] legitimate. or concern in particular, these entry level analysts and even data scientist jobs are gonna be replaced by ai.

That the idea of. A machine being able to write basic SQL or even heaven forbid, basic Python, you know, even going in and doing basic Excel formulas, you know, these are things that reasonably, AI can do or at least can do reasonably well. maybe this is all a terrible, terrible idea that, you know, the world of entry level data analyst and data scientists, even data scientists work is gonna be replaced by, you know, chat GPT or whatever.

or maybe it's clawed, but I don't know. What do you think?

‍

Avery: It is a really interesting. And obviously very important to, to think about and talk about. I wanna first and preface and say, I don't think actually anyone really knows what's a hundred percent going to happen.

and I, I don't either. But here's, here's [00:33:00] my 2 cents. one, I think. AI kind of existed before AI did in the data world.

like for example, one of my, one of my things I did as a consultant was I made like really smart web apps for, for companies. so a lot of the times we'd, we'd, we'd spin these up in Streamlet or we'd use like Dash from Plotly and create these web apps, for these small companies that did machine learning for them.

Displayed some data. It was like basically like, Hey, you do this repetitive analysis task, we automate it for you. Right. AI didn't exist back then, but we'd always start from a template an an existing template, right? And we'd copy and paste it and then we'd make all the changes. and that's kind of how I see AI working, where it's like, okay, I know I want to do this in general.

I'm gonna have AI do try to do it, and it's gonna get me 50% of the way there. So it, the way I look at it is more like, I don't know if you're a Mario Kart player, but like in Mario Kart, if, like, you can, like hit certain items and they will change how your, how your [00:34:00] car is going. And I see it like as the mushroom, it's the thing that makes you faster.

I don't see it as the thing that replaces you now will eventually, will it be good enough to replace people? Maybe. but I, think that there's so much untapped data problems and solutions at every existing company. Like I worked for ExxonMobil, you know, one of the biggest companies in the world. I can tell you that.

Like there were so many data projects we wanted to do, but like, we couldn't feasibly get to them all. I don't think it's gonna like lower head count. I think it's going to increase efficiency. that being said, also If you try to replace employees with this with ai, like I think there's gonna be a lot of issues.

I think there's gonna be a lot of problems. I use a lot of AI tools and a lot of times they suck. A lot of the times they're like, so like, I want to please you, that it'll do anything to try to do what I ask it to. So for example, and I'll try not to name names here, but like, there's certain tools I use that definitely are big players in the AI world.

I'll say, Hey. [00:35:00] Do regression with this. It looks at the data and it's like, I don't really know what to do. but I don't wanna disappoint Avery. So, I'm gonna create. A faux dataset and just do it on the faux dataset. And I'm not gonna tell 'em that it's on the faux dataset. And so like, obviously if that happens in production, you're kind of screwed.

And so the last thing I I'll kind of say is like I definitely see AI being used. I see it being used more as a tool. I think interesting analogies to kind of look at in other worlds, is radiologists Machine learning has been really good at detecting, you know, benign or cancerous tumors for years.

Avery: like we have robotics surgeons that still require surgeons. I think, I'm not like a pilot, but like, don't planes like kind of fly themselves now. And yet we have two pilots on every flight that we're on. Like that's kind of where I see.

The furthest that this data thing will go on is, is maybe we become more of like we do the beginning and the end and we kind of supervise the middle, but I don't see it replacing people personally.

‍

Anthony: No, I think that's fair. And I [00:36:00] think also to the earlier commentary, if the energy and focus of people who are trying to get into this space is on understanding, interpreting, presenting, And communicating results and information. Then the act of writing the code, writing the Excel formula, formatting the chart, like those are things which, if you can speed them up and use tooling to get them done, actually goes to improve your efficiency and, in a good way. your emphasis on the uniquely human piece that you can do, which is like, get the recipient of this to understand it, to make the business change. They need to make whatever the, the action that comes out of that, which speaks very much to where we started this conversation about how you think about portfolios and how you think about even, the work you're doing.

Like, don't jump into the tooling. 'cause maybe the tooling is going away. Maybe no one ever will write Python code again because. know, the, the machine will generate the code, but [00:37:00] communicating and interpreting, challenging the results. To your point about ma, ma making up data, like, you know, like probably a good idea to look and see what the code's doing.

'cause if it's generating a data set, then that's probably bad, right? And you should be savvy enough to understand that.

‍

Avery: A hundred percent. When, when we were at Exxon, we were building tools for, let's say more business side, of the business, right? And our tools would tell them what to do, what was optimal. There was still a lot of convincing. We had to do, like, this is why it's saying like the explainability behind it all, and like.

There's just, there's so much more to making a decision other than this is what the data says to do. We're gonna do that. That's, that's not how the majority of the world works.

‍

Anthony: Yeah, just watch the movie Moneyball and you'll be convinced that, it takes a lot of energy to convince an organization, to stop doing the wrong thing and start doing the right thing. well look, Avery, it's, it's been a pleasure and I, what I appreciate out of this conversation, is how, absolutely practical it is.

I think anyone either entering a career or [00:38:00] thinking about a, a career switch into the data space would be well served to, take your advice and think about the things you've shared on this podcast. So I really appreciate you making time and, and sharing your thoughts and expertise.

‍

Avery: Yeah, thanks for having me on. It was a pleasure.