The Data Hustle — Dr. Clair Sullivan

I was working on, you know, at the time we called signal processing, cuz the term data science didn't really exist yet. Um but, you know, it was a lot of mathematical analyses of things. So, I was doing a lot of research there. Um and some of my research was involving what is basically the mathematical predecessors to things like convolutional neural nets today. So, I was um so this was like wavelet analysis and stuff like that. It's like taking me in the way back machine, but um but that was just kind of my interest um my research interest in the labs and um I just kind of moved through the federal government journey for a little while, went um went to DC and did the fed thing

for a little bit. And then um with the administration change in 2012, um I was looking at, "Okay, what's what's the next thing?" And I decided I really kind of wanted to get back to academic roots um and I became a professor of nuclear engineering at the University of Illinois for a little bit. And that was when all of the, you know, the term big data hit and everybody needed big data and, you know, machine learning and stuff. And And so like data science, that started to, you know, really become just like general um conversation terminology. And so that was the work that I was doing there was uh with my graduate students and whatnot. Um but then I decided it was kind of time for

me to get out of academia um and a lot of companies were really hiring like mad for data scientists, machine learning engineers. So I took a job at GitHub as a machine learning engineer and got really interested in graphs while there. I was doing some graph stuff in the federal government, too, and that was a lot of fun. Um but then uh Neo4j had this uh developer relations role up and I was like, "Okay, what is that?" So I I emailed the guy who became my boss there and his I'll clean up the language a little bit on what he told me, but he said I I said, "What is this job?" He said, "Well, it's basically make cool stuff

director of data science and I was like well I got to apply for that because I'd actually started my first consultancy doing data science for the ski industry but the industry this was like 2015 the industry really wasn't ready yet. Um and so this job came up and I'm like well I have to apply for that and then I got it and did that for a little bit and then um corporate I I what I say um is that corporate quit on me in December of 23 and um the job market being what it was then and in what it unfortunately continues to be to this day, I said you know I'm the primary breadwinner, sometimes the sole breadwinner in my family and I

can't go without a paycheck so maybe this is the sign that I've been needing, the kick in the pants that I need to go start my own company and so I did. And um I've been doing that ever since. I've been very fortunate. I've beyond replaced my corporate salary. Actually I did that in the first week. Um and I know that's not typical but um but yeah I was I was very fortunate. I'm loving solopreneur life and I don't see myself as ever going back. >> Wow, it's that's Wow, I think you've hit most of the check boxes I guess. Working as a contributor, consultant, leadership, having your own organization. >> Yeah. Yeah, it's um you know I've I've tried a lot of

things. I've I'm solidly Gen X um, and you know, one of the things, you know, one of the kind of more depressing things about the job market is the ageism. And so, you know, when you're when you're looking at somebody who's been working, you know, since 2002, like a lot of people look at a resume like that and, you know, they don't they don't want to go there, you know, all these people are expensive, you can't train them to do new things cuz they're old and and I'm just, you know, I I've constantly learning new things, just absolutely constantly. So, um, it's it's much easier when you are as a solopreneur because people are hiring you for your expertise. They're not

>> Yeah, and this is something that I think

a lot of consultants get wrong or don't understand. And I didn't understand it, to be perfectly clear. Like this was one of my failings was um, you can have the best technical solution out there, doesn't mean people want it. People don't always want the best technical solution. When you're talking to people at that level, they have very different realities, and those realities involve things like, you know, I still have 5 years remaining on this tech stack. Or, yeah, it would be really great to build this thing, and it could deliver a lot of value, but I'm on the hook from the CEO to deliver something totally different. Um, and so, it when you're one of the things that I learned

being in those rooms was just how um, how political it is. It's not just technical. Even if they are like high-ranking data folk, you know, you could be talking to chief data officers. They They walk in a very different space, and they have to manage very different things. And the conversations that they're having really are not as technical as you might think. They may not even be technical people themselves. They They truly might may not. They could just be good leaders. Um, and so, when you as a consultant talk to people in in these levels, you have to understand their political reality. It's it Even if you can say I can drive this much ROI, or, you know, this will you you and you know, it's

going to make you this much money, or save you this much whatever, that's not necessarily something that they're going to need to use, want to use, be able to bite on. Um, and it's hard because like when I went into that role, I was like, oh, I've got so many great ideas about how to make this better, do that better, or whatever, and that wasn't what they wanted. Um, so, it's it's it's really it's it's a very different conversation that you have with those people. >> And that's like one of my favorite I think those for my career. One time I told somebody, "Hey, I don't think this solution is a good solution." >> Mhm. >> And uh, this person was also like a

C-level in the company, and what he told me kind of changed my career, which was "One, I don't care if this is a good solution. We promised to investors we were going to use this solution. So, it's not a matter if it's a good one or not, but how do we actually fit the solution into what the investors paid for? >> Yeah. >> And >> same thing play out and lots of money get wasted because somebody decided, "Nope, we're going to use this thing." And it you ask the question, "Why are we going to use this thing? Why this thing? You know, is this the best thing?" Sometimes those decisions are not made in the ways that you think they're made

and not always in the best technical ways and yeah, sometimes have to live with them and it's that's one of the not fun parts of it, for sure. >> Yeah, the the ones that I see often the parenthesis sometimes people just want a new solution because they think like the whole narrative of like, "Oh, we build it I build this stuff from the ground I deserve a promotion." Sounds great. >> Mhm. >> And sometimes even if you propose a solution that is amazing great ROI, they also don't like the narrative of I came and I stirred the pot. So, >> Mhm. >> that's very related to what you're saying. >> I think Gen AI is a great example of that, right? You know, and this is you

And that can be a totally fine answer. Nothing is a fine answer. Um so, they go and they get Gen AI, you know, they do they spend a whole, you know, crap ton of money on it and it's then they get mad because they're like, "Well, where's Where's the value I'm getting from this? What's What's it driving in the bottom line?" And the answer is, you didn't go in starting with that presumption. You came in saying, "Well, I have FOMO. Therefore, I need GenAI." And that's the wrong way, in my opinion. >> Interesting. So, I guess the conclusion here is sometimes when we're talking to C-level people, we don't have to think about what we propose so logically, but more

from a political lens. >> Yeah, I mean, you know, the the the nature where they're at, and you know, it's really important that consultants try to meet them where they're at in so much as possible. And where they're at is that they are dealing with political realities that the rank and file does not see. Um and won't necessarily make sense even. You know, the a CFO looks at a company entirely differently than a data scientist does, or even a senior manager. There's this, you know, as you start going up within the the layers of management, director is kind of an interesting position because like senior managers, their job is to manage their people below them. VPs, their job is to manage the people

>> Nope. I wasn't. I'll be perfectly honest, I wasn't. Um I What One of my uh

you know, just one of my personal brands is honesty. And I I'm not going to say I was asked to lie, but what um people who work for me, who come to me and say, "Hey, we would we would like to work with you." They know that because I'm not going to you know, I'm not somebody who sugarcoats a thing or just is like, you know, going to say, "Oh, hey, I think this is a great idea. Let's go ahead and do it." You know, that's you know, senior boss person like, "Yeah, let's you know, if if I see something that I think is um is going to not result in what you think is going to result in, I will call that out. And not

everybody wants to hear that to be perfectly frank. Um and so the people who work with me, who I have long-term relationships, they do value that input. And so it's kind of it's it's like from for a consultant's perspective, like when you hire a consultant you should want that, right? You should you're you're paying a lot of money to a person who's coming in and evaluating a thing for you or you know, maybe they're freelancing and they're writing code for you or something like this. But like you're you're not managing that person. You're you're bringing them in because they are an outside set of eyes. And but still not everybody has fully embraced what that means, that that's they're going to not necessarily tell

>> Interesting. Um so, first off, there's no one right way to do it. Like, and especially if you're doing this solo, you're you're you know, you're your own boss, your own company of one. Um one of the great things is that you can discover what does and doesn't work and pivot immediately. You don't have to ask permission to do it. Um so, if I was a wildly technical person, like, you know, we typically think of consultancy as somebody who comes in and evaluates something and their work product are are reports, slide decks, things like that. Um so, if that's what you like, great.

Do it. Um and people who have spent, you know, a few years doing technical things, they've seen things go well and they've seen things go poorly. And the more you do it, the more you start identifying the patterns in that. So, you know, you as a technical person can identify those things. So, I personally would say focus on like technical the the technical side of the consultancy, you know, not necessarily the political side or the business side, strategy, things like that. Um versus if you've had some time in management, then you can perhaps lean in if if you're interested. You don't have to. You can do whatever you want. But you could lean a little more into strategy or something

Do you want it to be code? Um because a lot of times we'll call that freelancing. Um if you want it to be reports and analyses, you know, that's consultancy. You could do product-based. Um you could do, you know, training, which is something I'm leaning into more these days cuz I just really like teaching. Um but I would say focus on what is the what is the thing that will get you out

of bed in the morning because it is a hard it is a hard path. It's not there there's you're going to be doing bits of work that you never thought you'd be doing, you know, you're just because you are doing data consulting, you know, you maybe get to do that like 60% of the time. You know, you've got a lot of other work that you have to do that are not necessarily part of the skill set that you've been trained with like business development, accounting like oh dear God, what I have learned about accounting in the past few years. Um stuff I never really wanted to know, but you know, it's kind of like the CD side. Yeah. Um you know, and then you know, another

thing I would say is like you know, I bring up accounting, but understand where your skills are, what you're willing to spend that non-data consultancy time doing and what you're not and and pay for somebody to do the ones that you're not willing to learn or capable of learning or have time to learning you know, I pay for an accountant for a reason. Um I still have accounting stuff that I have to do like bookkeeping to prepare stuff for them, but um marketing. Oh my gosh. Like oh my gosh, I am not a marketer. Um I don't understand anything about marketing. I'm taking online classes on marketing. It's it's a slow if I had enough money, I'd be higher Okay, so I'm

putting this out there. Anybody who wants to do cheap marketing, give me a call, you know, like Um but that's that's hard. Business development is hard. Um you know, establishing relationships with potential clients and, you know, reaching out to them in ways that are effective. Um you know, those are not things that they teach you when you go to school for, you know, computer science. >> So, it's really about like learning what to delegate. >> What to delegate? Some of the stuff you can't. Like, you know, especially like for a solopreneur whose business, you know, is based on do people know me? Do they Do they value what it is that I provide? I have to still be the face of

the business. So, like >> Personal brand. >> Yeah, it's personal brand. Um so, I would I would look at where are the places you can delegate and where are the places you can't and know that you're going to be spending some quality time in learn a very steep learning curve. >> Be ready to do all of them, but hopefully when things start going well, you'll just delegate the parts you don't feel like learning. >> Yeah. Well, and then like, you know, hopefully you go off and you do this and you're wildly successful and you wind up with having the good problem of I took on more work than I have time to do. You know, and then it really helps that

you know other people doing this. Because like other consultants and solopreneurs, these are not your competition. These are your colleagues and we pass each other business all the time. So, it matters that you establish a network of other freelancers and solopreneurs and consultants and whatnot. So, that way somebody comes to you and says, "Hey, I'd really like to enlist your services." Like and you can say, "You know what? I'm fully committed at this point, but I know this person over here. You gain credibility by just not taking on every single bit of work. Um being able to refer out. Um and other, you know, other solos will do that with you as well. So, >> Yeah. Well, that's an interesting point.

>> a great question. Um You know, I'm not going to say that I 100% know the answer. Um but I can tell you that there are like you start working around the solopreneur world and you'll figure out the handful of people that are like you know, like so I work in the graph space and I know that if I am over subscribed, just because I had a beer with somebody, no, but I'll know other people who know

that person. Um so, you know, life is a graph and it's all, you know, about the connections that you make in it. And so, yeah, I I wasn't paid to say that. Um but yeah, so it's like, you know, there are people that I've known in the graph domain who are solopreneurs, freelancers and I've known them for years and I wouldn't hesitate to send somebody to them. Um I know other people that I've worked with in the past who are out and starting their own solo thing and it's, you know, so I've gotten some experience talking with them and it's, you know, it's more than you can't just have the one beer and call it a strong connection in your graph. Like you have to have

And that's that's something that I teach in my solopreneur course is a it was like like these lots of little tricks that I learned in DevRel to quickly grow community and establish

You're going to start making tweets about it or if if you're on X or Twitter. I I'm not right now, but um and everything, you like you pick like one little snippet, the micro content out of the macro content and then you can start generating content around that. And everything points back to the macro content, but like, oh hey, you know, I learned to do this thing while writing

this blog post like here. Um and let me talk just a little bit about this thing. Go check out the big post. And then there can be a video with it and the LinkedIn post to go with it and they all point back to each other. So, one bit of macro content generates hundreds of bits of micro content. Um and it's it's just a flywheel that you I call it the content flywheel in my course. And you just keep rinse and repeat. >> Well, I guess we can make a lot of shorts out of this interview. I'm already >> Indeed. >> Yes. >> We'll put your advice we'll we'll we'll put your advice into practice after this talk for sure.

>> Ooh, that's a good one. Um so, you know, when I was first starting my business, I was I I'm actually not working on what I thought I was going to work on. Um I thought I would be a consultant to companies on data culture. Okay, and I was super excited about that. I still am. I'm still very passionate about that subject, but frankly nobody wanted it. Um everybody wanted GenAI. So, I was like,

all right, give the masses what they want. But like my technical background prior to the GenAI revolution graphs and natural language processing. That was another thing I did a fair bit of. And like those are all the things that just naturally led themselves well to GenAI. Um Now, you could be somebody who's like, you know, working in data, is not really interested in GenAI, even though it's like what the whole planet wants. I would say instead of framing it as the technology, framing it as what problems are you solving? So, um you know, one of the things, you know, hearkening back to my PhD and going to work, you know, in nuclear emergency response, um was you know, what what are the what are the

subject areas I'm passionate about? So, I, you know, for lack of better term, I call it do-gooder stuff, meaning I like to do things that contribute a contribute in a positive way to humanity. So, like some of my clients with GenAI, you know, and I have lots of friends who really hate GenAI, think it's the you know, the most evil thing to be unleashed upon humanity. And you know, and I point out, okay, but my clients are doing things like looking at food insecurity in Africa. Or migrant um migration out of war zones. Um money laundering amongst oligarchs. People trafficking, things like that. And using GenAI to go after those things. So, you know, do I like GenAI? Do I hate GenAI? I don't know. It's like

saying I like or hate the internet. But um I'm solving problems that I'm passionate about with it. So, instead of thinking of as a technological thing, think of it as a problem domain thing. >> Interesting. Yeah, I've I've heard people already talking about GenAI as the new plastic. >> Mhm. >> So, it's not a bad thing, but it's like the mindless over consumption. >> Mhm. Yeah. Yeah. It's going to be interesting because like, you know, these are the same arguments because I was around, you you at the start of the modern internet. I'm not going to say ARPANET or anything like that, but but like, you know, I was part of the group of kids who was running bulletin boards out

of their computer in their house with dial-up modems. I remember dialing into CompuServe at 300 baud, not kilobaud, but baud, because you paid by the baud rate, and 300 is about as fast as you can read text scrolling across the screen. You know, so like I remember all of this. I remember AOL and and all of that. And I remember a lot of the same arguments being made then as they are today. And there's there's a there's a psychological and an ethical evolution that occurred then and will occur now, but we're just at the start of it. It's this these are still early days. >> Wow. Well, Claire, actually, jumping a bit of topics, maybe a bit abruptly, one thing that I really want to ask you

>> Yeah. Yeah, okay. So, this is this is a great question. Like, um I think it's one that people are really scared of because they don't know how much things cost. Um so, there's two things I recommend people do. First is and and this is going to be a painful exercise. I warn you right now. It's demoralizing.

Figure out what your hourly rate is. So, what you do is you take your your annual salary, and in the United States, you divide by the number of working hours in a year. In the United States, that's 2,080. So, that's like 40 hours a week, 52 weeks a year, assuming you work 40 hours a week cuz, you know, most people work more than that, but that's just the math. Okay, and you're going to get an hourly rate by doing that. In In UK, it's different. Like, you want to look up what this number is for your country. And um from there, you get this hourly rate, and you're going to do it, and it's really low. Like when you compare

it to like what your hair stylist makes, like what you're paying them. And I'm not saying it's what they make, it's just what you're paying them for an hour of their time, or a massage therapist, or something like Like it's low. It looks really bad. Okay, that's not That's your starting point number. That's not the number you that you should charge, because you're going to have expenses when you go out on your own that you don't have working in corporate. You have to do things like pay for health insurance, and in the United States, that's an incredibly painful number that you're looking at. Um you need to you know, be contributing money to retirement. These are things that when you're working in corporate,

frequently your company helps with. And now you're going to be doing that yourself. So that's your starting point number, but then you want to, you know, think about how do we create a burdened number here? And the rule of thumb that I was told was take your salary, multiply it by 1.25, and lop off the last three zeros, or divide by 1,000. And that's a good starting point for your hourly rate. Now, a lot of people say, "Well, I don't want to work hourly." And that's there's a lot of reasons to not want to work hourly. A lot of places um like if you're working in the with the US federal government, they require you to. Like there's there's policies in place that it they

will pay hourly. That ends. It's not project-based or retainer-based or anything like that, but it's it's just a starting point for you that will then help you price out things like project-based work. So if you know what that number is, you can then say, "Well, you know, let let's I'm just going to make up a number. Let's say it's 150 bucks an hour." Um and let's say you know, well, this is going to take me, you know, this client wants this thing. It's going to take me 100 hours to complete. Okay, well, that now you know, then if you are going to just break even on your hourly rate, then you just multiply those things together and that's that's how you can

get your pricing. You know, and that assumes things like you know, there's no travel, there's no infrastructure costs, things like that. But you know, this is kind of like a good starting point. A lot of people don't like hourly because it's you know, it's kind of incenting the wrong behavior. But it's just a good number to have in your head. And I also encourage people to have that number in their head even if they're staying in corporate. Because a lot of us have found ourselves on the receiving end of these conversations with HR lately. You know, sorry your role has been eliminated. Okay, just because they eliminated your role or 10,000 roles doesn't mean they have 10,000 people less work to do. It means they're trying

the hourly rates were for contractors and companies that I've worked with. Um and I figured out my hourly rate sort of with these same approaches. I have some other tools that I give people in my classes. We go through the what I call the financial readiness framework where you look at how many hours at what hourly rate and what expenses and taxes and blah blah blah do you really need to make. And what I found was it was a fraction of what the big contracting companies were charging. You know, I talked to a friend who worked at I'm going to say a Fang company. I'm not going to name the company, but they offered professional services. And I said, all right, well,

if I was you know, if I was a company getting professional services from you, how much is your company taking off the top as profit? And he said, you know, it's roughly a third. Okay, that money is not winding up in the hands of the contractor. So automatically, I know that I am cheaper than working now it do I have the backing of a Fang company? No, or you know, like there's not the name associated with that Fang company if you're working with me. However, it's just kind of like that ballpark idea. And what I found was that yeah, I was coming in you know, at half 2/3 what the big you know, contracting companies were charging. So So that was pretty validating to hear

Because when I compare prices, I see some people that charge as little as a couple hundred per day. >> Mhm. >> Most of them remotely from another cheaper region. >> Sure. >> But then there are also like very, very expensive ones and you see people that you might think are better than you charging less. >> Yeah. >> that's very challenging. >> Yeah, two thoughts on that. First, the place where I see that the most is on online platforms like Upwork. Um Upwork is a race to the bottom. And um you know, so you'll see you know, work that's put out there that I want a

full stack GenAI solution fully prod LLM of owls blah blah blah and I'm going to pay 20 bucks total for the whole thing. Okay. No. Um now and there are countries out there with significantly lower cost of living than say the United States where I live. And you know, that can be you know, very you know, that that can be a lot of money for these other countries. I avoid those platforms specifically because of the race to the bottom. Um then I also look at um the clients that I've worked with, clients who've talked to me when we've talked about, say, hourly work or project-based work. Um the clients who are looking to pay those bottom rates are not clients you want to

work with anyway. Um there it's it's a sign. It's a red flag. They don't know how to work with a contractor. They don't know the amount of work involved. They don't value it. Like, they should know what's the market rate. And if they don't, like, you can you can help educate them on these things. Like, I'm able to say, you know, if you were to go to you know, one of these big contract shops and pay for somebody to do the same thing, um this is what they're going to charge you. So, you're actually, you know, getting a better price with me. Or, you can, you know, you can go with a a cheaper country or something like that. And, you know, in

Like, have you seen that change in the last few years? >> And I have clients who have had this very conversation with me. Um

my thing is, what is it that you are paying me to do? Okay, I I have a teenage kid who could code with GenAI. She hates coding, but, you know, if somebody came to her and said, "Write me software to do XYZ and I'll pay you this amount to do it." She could do it. Anybody could do it. Um but, what people are paying me for isn't that. They're paying me for my expertise. Okay. So, yeah, I could go vibe code something. Cool. But, what what you have to If you're going to successfully create something with vibe coding, or if you're going to successfully create a report, anything that you're going to create with GenAI, you still have to come at it with enough

expertise to be able to say, "This is exactly what I want. This is why I want it. No, that's a hallucination. No, the client isn't going to care about this thing. They want more of that thing." You know, so you still have to be able to prompt GenAI properly to create the right thing. And I've certainly, you know, created stuff with GenAI that, you know, really looks great, but my expertise knows that it's garbage. Okay? And that's not it. I You know, I create stuff with GenAI all the time. Not all of it is garbage. But, um at the end of the day, it's a question of you're not hiring a you know a mechanical Turk here. You're hiring this thing that's got these gray

somebody to really put some thought and develop something that's never been developed before. So you know, I might vibe code a little bit of it. You know, hey I like I will never ever ever remember regex ever again. Like ever. I'm just I remember the very first time I used Copilot. I was a beta tester for Copilot and that was the first thing was "Huzzah! I never have to remember regex again." You know, so >> Thank god. >> You know, so like I there are efficiencies that I can get that are you know less consequential by using GenAI in writing code for example and I know where they are and I know where there are places where I'm like I just don't you know, I'm still going

>> I think if if you have played with AI and tried to do your work with it for at least one like 10 hours, you're already aware of many of these issues. >> Mhm. >> But it's more than it's also part of your job now to educate people when it's still a good idea to use your judgment and what things can be easily automated. >> I The analogy that I give is it's like when I was a professor. Okay, so I had an army of students and, you know, the PhD students in particular who I'm thinking about right now. And I, you know, my job was to guide and lead their research. Um, but ultimately, you know, they were smarter than I was

and that, you know, whether they know that, I hope some of them are listening right now, but they were smarter than I was and my job was just to ask questions to them and get them to think and then see stuff come back and say, "That doesn't look right. Why don't you go redo that? Try it this way." You know, I wasn't the one coding it though. They were the ones coding it. They were the ones doing the work. I was just asking the intelligent questions about it, um, and, you know, flagging it when it didn't quite look right. So, you know, to me GenAI is the same thing. GenAI is the grad student who you go and send off, "Go do your research, you know."

right now if anyone is interested to uh follow you along? >> We'll leave it in the description here. >> Absolutely. So, you know, my webpage clairesullivan.com and my parents, you know, played a joke on me in the spelling of my name. My name does not have an e, so it's just c l a i r sullivan s u l l i v a n dot com. And um what I'm working on these days, I'm working on growing out the education side of my business. So, I am doing training um GenAI for non-technical people, technical people, you know, very hands-on, very not hands-on, very customizable. Um I will be doing some upcoming courses on GenAI for non-technical people, for office workers who need to just get

efficiencies out of it. So, um that will be uh hosted probably on Maven. Um so you can find me on Maven, but my website also will, you know, have links to all of this stuff. >> Excellent. We'll make sure that everyone can find all these resources in the description and uh, I would like to thank you again for joining us today. I had a wonderful conversation here. I don't know about you, but I learned a bunch here. Uh, we're very happy with this conversation and uh, yeah, hope to speak to you soon. >> Great. Thank you so much. Really had a great time. >> Thank you for your >> All right.