Disclaimer: The views expressed by our interviewees are their own and do not necessarily reflect the views of GradSimple. We only edit these responses for clarity. All opinions remain those of the interviewee.
- Year Of Graduation (Bachelor’s): 2018
- Year Of Graduation (PhD): 2024
- College/University: Cleveland State University (undergrad), Colorado State University (grad)
- Undergrad Major: Mathematics (Physics minor)
- PhD: Mathematics
What do you do for a living, or how are you supporting yourself right now? Can you share what a typical day in your life/at work looks like?
My main project at the moment is working with AI models on what’s referred to as stumping. It wasn’t very long ago that a recruiter on LinkedIn reached out to me for a company called DataAnnotation. They have clients that feed them AI models, and they want to get information on how well those AI models do at solving certain types of problems. Essentially, I am given access to the AI model and my goal is to feed it a problem that it will not answer correctly and then write a report about how it failed, why it failed, and what I think should be done differently for the model.
So, more or less, I’m getting paid to make AI models do badly at math. Eventually, me and DataAnnotation parted ways, but another company called Outlier came along that works for me a lot better. I’ve been doing work with Outlier for the last nine months, stumping their clients’ AI models on a whole variety of different mathematical tasks.
A typical day will absolutely involve a lot of, you know, stumping. That’s how I make my money. Besides that, I also spend a lot of time looking for jobs. I apply for lots of stuff every day on LinkedIn and Indeed. It’s not really gone very far, but you know, that’s the grind that we all face at this point in our lives. I also spend a bit of time trying to work on personal projects or learn additional skills, which is where I feel like I’m most disorganized in my life.
What I work on from one week to the next varies quite dramatically. One week I designed a private app for my phone that is meant to track my diabetic health, since I’m a type 1 diabetic. I’ve done projects to build web scrapers to get a bunch of data about places in the US, so I could see, if I got a job there, what areas might be more affordable than others. I’ve tried to learn a bunch of probability theory since I know a lot of different job descriptions ask for it in my field. I’ve also been trying to work on a publication based on my dissertation for the last year, but that has been sort of slow going because as important as it is for a person to publish in my world, at the same time trying to keep my head above water financially is more important.
Do you think AI stumping is a viable career path? Is this primarily contract-based, or are you freelance?
There are definitely some companies that do contract work. That’s not unheard of, but most of them, including the one I work for, are more freelance-based, and it’s easier to get into that. It is not a constant stream of work. There are weeks where I am given opportunities to work basically as much as I want, as long as I want, and make about as much as I want to make, but there are other weeks where there will be nothing.
I’m currently going through a dry spell where I haven’t had anything to do in the last two weeks. That does make for a somewhat concerning outlook. I’m not really in a position where I can afford to not take what comes my way. At the end of the day, I would say realistically, someone should probably not look at this as a career so much as it’s just a good way to make some extra money. From my personal perspective, I feel like this is something that is a “unique moment in time” kind of thing. I think in another three or four years, these kinds of jobs are probably going to mostly dry up as the big AI boom starts to taper off. The companies that wanted to train their models will have gotten what they’ve wanted, and other companies won’t find it viable or profitable to keep pushing it. So, I would honestly describe myself as very lucky to be here at this moment when I most need that work.
Taking on this kind of part-time work can be a good way into a company as a permanent hire. Is that something you’re hoping for? Do you think that’s a possibility?
It’s not impossible, and I would certainly be open to such an opportunity. However, my understanding is that most of these companies are very selective about it. They don’t really have a place to apply, but rather if they like your performance, they will reach out and give you an interview. Usually though, I would also say that because they’re a third party as opposed to the company that actually builds the AI, these more senior positions are people management roles. I’m certainly down to throw my hat in the ring if they want me to do that.
I get the impression that, in general, they expect the mass majority of people who flow through their company to be there for maybe a month or two, stump some models for a little while, and then move on.
What is your position on AI? I’m curious if you perceive it as a threat, or as an important advancement, or somewhere in between.
That’s a good question. I want to preface this by saying that I knew many people in graduate school who did PhDs on topics that are directly related to AI, and they spent a lot of time thinking deeply about AI procedures whereas I most definitely did not. There are other people who probably have a more refined opinion of this. However, for myself, I definitely will say that at some level, it often feels like training my replacement. These AIs do surprisingly well. When I’m asking them questions to stump them, it’s not like back when I was a graduate student and teaching classes where the questions I asked were meant to help the students learn. No, the questions I’m asking the AI are intentionally some of the most ridiculous questions imaginable simply because I’m trying to find the things that the AI’s data set was not trained on, which often are things no human would ever want to ask in the first place.
As a result of this, I would say that I am a little bit concerned in some regards in a lot of tech spaces and places where mathematicians with my kind of background are utilized. It definitely feels like this is somewhat of an existential threat, but it also probably isn’t an existential threat that will manifest in the short term.
Why did you choose your undergrad major?
I was always someone who was excited and invested in mathematics. I did well at it in high school and it seemed like an interesting direction. I also remember discovering the YouTube channel Numberphile, which is hosted by Brady Haran and he has guests on every week talking about their favorite mathematical concepts. For teenage me, that was exciting. It was cool to get to see all these unique things.
When I entered undergrad, I had this idea that I was going to become a college professor of mathematics, so I started pursuing a bachelor’s degree in math. Along the way, I had fun poking my nose into physics a little, and I got a minor in that.
Ultimately, I ended up falling in love with the more theoretical side of math. I got my bachelor’s in 2018 and wanted to go to grad school to continue pursuing my dreams. But one thing that is important to recognize is oftentimes our expectations of what things are end up being quite different from the reality, especially when you don’t have a lot to go off of. I’m a first generation college student, so none of my family or family friends were in that world, and I didn’t really know what being a professor entailed, or what true academia was, or what grad school was, and things like that.
Can you tell me about your experience after graduating from undergrad? Are you happy with where your journey has led you thus far?
My goal was to immediately go to graduate school. I wasn’t that interested in trying to find industry work, so I started applying in 2017 before I graduated. I knew I was coming from a smaller state school, and I didn’t really expect high-end school acceptance, but ultimately I was told to apply to 10 schools. Three should be sure things, where you’re very likely to get in. Three should be long-shot schools you’d love to attend, and four of them should be middle of the road.
Ironically, I ended up getting into two places, both of which were in that middle category. None of my backup schools actually accepted me, which was sort of funny. One of them was here in Ohio, and after spending a lot of my life in Ohio, I wanted to do something different. So I chose the other one, Colorado State University, and got to experience a very different environment. I went with the intention of getting a PhD in mathematics. I did pick up a master’s along the way in 2020, but that was absolutely my goal.
There was never really any break between undergrad and graduate school for me. I got very lucky, and I am definitely grateful for that because there are a lot of other people who come from a background like mine going to a smaller state school where getting into a good PhD program is a bit of a long shot.
How did your graduate program work, between your Master’s and PhD?
At Colorado State, they essentially have what they refer to as part B. A lot of people who do PhDs are probably familiar with the idea of having to do qualifying exams. Part A is essentially doing your qualifying exams, but in order to be considered a PhD candidate, you also have to finish what they call part B, which is usually a smaller project to demonstrate your research fortitude and capability.
My advisor believed that because a master’s is essentially interchangeable with that anyway, I might as well get a master’s so that I at least have something I can take away from the experience, just in case. That seemed reasonable to me. This also happened in early 2020, so I didn’t have a lot else going on, like most people. I think that for a lot of people, 2020 was a year in which you basically just kept doing whatever was on your to-do list and tried not to think too much about the state of the world.
How would you describe your grad school experience, now that you’ve completed it?
I would say that grad school is an interesting place. For the most part, I enjoyed it, and I think I learned a lot of valuable lessons both as a mathematician but also as a person. My general takeaway from having been there for the years that I was, was that academia probably is not the direction that I want to spend the rest of my life. There are things about the way that it is run and the bureaucracy that aren’t super appealing to me. I think most importantly, I learned that I’m the kind of person who needs a well-established work-life balance. You really don’t get that in academia. My advisor, for example, would sometimes answer emails at like 2 a.m. because that was the most opportune moment for her to do so. I think I realized that really wasn’t the life I wanted to live forever. Graduate school was a very good stepping stone for me, and I’m very grateful I did it, but I think six years in that lifestyle was enough time.
I started to have a lot of those realizations in late 2019. I had a really challenging course that fall, and it was one of those classes that most graduate students will have where it ends up eating most of your time. I walked out of that semester exhausted, and then COVID happened, and suddenly the entire world shut down. It was a bit of a shock to the system. I think I realized that the way I had been doing my work up to that point was unsustainable. I’d been working like 80 hour weeks to get through everything that semester, and it was just too much.
I think I learned several things. One was that my work-life balance was important. I needed to better take care of my mental health. The second was that after spending a lot of time thinking about deeply esoteric things that had very little connection to the real world, I really wanted to be working on problems that were a bit more connected to reality, which led me more toward cryptography. The third was that academia was taking a bit of a toll on me and that I probably needed to think about other options.
Could you explain the general idea behind a cryptosystem, and how they work?
Cryptosystems are algorithms, usually based heavily on a mathematical process, that allow two people to share information safely without a third party interfering. This could be as simple as exchanging messages, but almost every security system online is based on these principles. Mathematicians and computer scientists like to refer to the two people exchanging info as “Alice” and “Bob.” Usually a person trying to intercept their secret information is called “Eve.” Alice and Bob both start off with a piece of information, a personal secret that they don’t tell the rest of the world. There’s also another piece of information everyone usually has access to.
Alice and Bob use their secret information on the public information, kind of like a “key.” They then hand each-other the result and do it again. These algorithms are carefully chosen so that using Alice’s key and then Bob’s key is the same as using Bob’s key and then Alice’s key. They end up with the same end result, which can be used to construct a secure connection for trading secrets. Eve doesn’t know the end result, and the problem is chosen to be sufficiently hard so that she can’t reverse engineer anything from public information.
I studied a post-quantum cryptosystem that works as follows. You have a network graph (a bunch of points and edges between them). Each point represents a mathematical “shape” and each edge is a relationship that exists between these shapes. There’s a starting shape that is public knowledge to everyone. Alice and Bob go for a distinct walk on the graph. Something like “Left, Right, Middle, Left, Left, Middle, Right, etc.” They end up at a new shape, and then share that shape with each other. They then take the same walk again, from the shape the other stopped at. Because of special mathematical properties, they end up at the same place and can establish a secure connection. The graph is unimaginably large, think billions of nodes. Because of this, Eve has no way of reverse engineering what path Alice or Bob took. As such, she can’t find the end point they arrived at. This is a bit of an oversimplification of course, but it’s the basic idea.
Can you talk about your PhD thesis, and what you worked on?
In 2020, I was getting more and more interested in doing something that had applications to the real world. My advisor started guiding me toward cryptographic problems, and for my master’s I’d been studying an object called the Prym variety, which is directly tied to these other things called Abelian varieties. It just so happens that Abelian varieties are highly relevant to one specific cryptosystem that was a candidate, at least at that point in time, to be the post-quantum cryptographic protocol: something that would be secure from attacks by quantum computers.
In the 1990s, a guy named Shor invented this algorithm where if you have a quantum computer, it breaks everything we currently have for computer security. Like things that are supposed to take thousands of years for a normal computer to crack are now broken in seconds. It’s really bad for anyone who cares about, you know, actually having security online or governments maintaining their secrets. So the National Institute for Standards and Technology, NIST, sent out a call to mathematicians, since mathematicians typically are the ones who build good protocols. By the time I entered the scene, mathematicians had come up with four different models.
One of these models was called Supersingular Isogeny Diffie–Hellman, or SIDH for short. My area of expertise with these varieties essentially is very adjacent to it. I became interested in exploring a higher-dimensional version of what researchers were already studying, so instead of using one-dimensional objects, I focused on two-dimensional ones. People before me had more or less concluded that using two-dimensional objects was too slow, however. Even if you could get the cryptosystem to work, it wouldn’t be efficient enough to be worthwhile.
It’s kind of ironic though, because when I started to build expertise in this area, in 2022, something absolutely wild happened. Two mathematicians in Europe managed to break this proposed cryptosystem that used one-dimensional objects, using an attack involving the two-dimensional object version. Suddenly my area became really interesting to a lot of people because it’s now like, “Okay, well how far could you take this attack? How viable is it? Are there ways that we can patch it up so it still works?” Suddenly, for a very brief period, what I was doing was a hot topic.
For my thesis, I wanted to first build code that actively calculated these graphs because most people had only done theoretical work, like, “If we have these graphs, what would they look like?” I wanted to spend time actually building them. The second thing was then asking questions like, “If I have these graphs, what can I say about small loops in the graph? Can you take walks where you start at one node, walk around the edges and end up right back where you started again without walking the same edge twice? How common are those? If you start off with different kinds of nodes, do you see patterns arise? Are there commonalities?” That became a lot of the kinds of things I was asking in my dissertation for these graphs, more built on the two-dimensional objects.
What does purpose mean to you? Do you think you’re living a purpose-driven life currently? Why or why not?
I think purpose for me comes largely from doing things to help people, whilst still being interested in what I’m doing. I’m the kind of person where my goal in life is to find ways to take my passions and create good results for other people. I don’t feel like I’m living this purpose at the moment, sadly. I feel like I have yet to really find my niche in the world. We were talking earlier about my work with AI companies. I think that in a lot of cases there is some importance to what I’m doing. However, it’s hard to argue that most of the tasks I finish are helping anyone or bettering the world in any way. Most of the training data probably won’t even end up being used.
What’s your biggest struggle right now? What led to it? How are you moving forward?
I would say one of my biggest struggles in life is that I tend to be somewhat of an anxious person. I’m a bit of a perfectionist. I’m the kind of person who really wants to see myself succeed. I try to create a lot of plans and think of every possibility. It’s very easy sometimes to just worry that if one thing goes wrong, all this other stuff will go wrong, too. During COVID, it really became difficult with watching the whole world slowly falling apart, and I had to learn that sometimes, things were just going to be good enough. I had to do a lot of meditation and journaling to actually get to the place where I felt comfortable with that. I learned some breathing techniques online. There are a lot of really good resources for these sorts of things.
At the end of the day, I think that I just had to rework the way that I thought about things. I still stress more than I’d like. I think that stress is just a natural state of being for me. But I’ve gotten better at finding ways to deal with that stress, which is good. I don’t think myself five years ago would have been able to deal with my current situation.
What do you want to spend more time working on, either personally or professionally? What is stopping you?
In my life, I sort of see a triad of things that I personally value as important. I want to be working on problems that I find interesting and that speak to me. I’m perfectly fine with that coming from my work. I think that’s quite valuable and I still do that a bit when I work on my personal research, but a lot of my time gets eaten up by the AI work or applying to jobs, or trying to build skills. A lot of PhDs are very fixated on working specifically in their niche field. For me, as long as it’s something that is interesting and it makes me curious, I’m happy to do most things. I just want something to pursue that I think is fascinating.
The second piece of the triad comes from something that I definitely feel a lot more limited than I would like it to be these days, and that’s travel. One of the things I really loved about graduate school was the opportunity to go to conferences. I have a great fascination with this world, and I want to see more of it. Part of the money I’d like to earn in my lifetime will go towards travel.
The third thing I’d say is just to have the time to pursue hobbies. I am a bit of a violinist. I like cycling. I like cooking things. I’m very fascinated by trying new recipes. I’m a bit of a gamer as well. I would like to have time to put towards those things. I think that realizing that about myself is part of the reason why I wanted to move away from academia because, unless you are very good at managing your time commitments, it’s much harder to have passions outside of your field of expertise.
How do you define being a successful person?
I think success in my eyes comes from reaching a point in your life where you have the freedom to pursue your interests without worry that it will interfere with your life. I guess that’s a simplification, but I feel like there’s a lot of different things that people value. At the end of the day, I think it comes down largely to freedom, and I guess if I had to paint a picture of what that looks like, I think that varies from person to person. Some level of financial stability is very important, but I do think that from individual to individual, success can differ.
What’s the biggest lesson or piece of advice you wish you’d known earlier?
I’m going to give two pieces of advice that I think are important. The first is simple: always have something you are doing besides the thing that you do for school or the thing that you do for a job. Have a passion outside of your work, and that should arguably not be a side hustle either. Something that brings you joy, that you can spend time on and commit to. A lot of things in life, especially graduate school, can be very soul-sucking. It is important not to make that your entire personality.
The other piece of advice is if there’s something that you value and you want to try, just go for it. I feel like a lot of people get really caught up in, “Oh, this thing will make me the most money. This thing is the most reasonable path forward.” I’ve seen plenty of people who went out of their way to do something that they thought would make them a lot of money who are now out of work and whose prospects for making a lot of money are non-existent.
Connect With Vlad:
- LinkedIn: @VladSworski
- Website: VladSworski
- Link to his dissertation
P.S. Wanna see more personal interviews? Subscribe to GradSimple and get the good stuff delivered straight to your inbox, once a week!
