George Mendoza is a Data Science Career Track graduate currently working at an artificial intelligence consulting startup based in Texas. We asked him to reflect upon his Springboard learning experience, including his capstone projects, which played an integral role in his finding a data science job. You can read more about the projects on his website and peruse his LinkedIn profile for further context.
I have an addictive personality—or something resembling it. In the process of understanding this, I’ve turned away from suppression and toward deflection. If I’m gonna chase a dragon, it’d better be something productive. So, after a long and critical dialogue with myself, I decided my first independent data science project, my introduction into this new world, had to be on baseball gambling.
Turned out to be a pretty minor deflection.
As part of my Springboard capstone work, I found a comprehensive baseball database online—released annually with data from the previous season—and stumbled upon a sheet of gambling odds going back nine years. The goal: use predictive analysis to find a profitable mix of expensive favorites and high-value underdogs. The best-performing models for each category made a combined 350,000 units of what I called “alleged U.S. dollars” over eight seasons (when wagering 100-500 units at a time), but showed impressive returns as I re-trained my model with each year’s data.
As it turned out, the Supreme Court had an inspired moment of wisdom and transformed my project into a legitimate potential source of passive income. As a new fountain of money intoxicates sports leagues, the data surrounding them will only become more and more accurate. And data provides the greatest opportunity to bring some level of objectivity to this sordid business.
The obstacles in front of me as I consider turning this into a real business: taking this theoretical approach and applying it to real-time games requires a new web scraping element and a potential data set change, given the limitations of what’s readily available. The bright side: my models weren’t that complicated. I now have an opportunity to revisit the project and apply the techniques and approaches I’m newly familiar with, or use the vast machine learning resources online to learn some others. I consider this project, at the very least, a great proof-of-concept that Vegas is about to lose a lot of money to me.
Investing in stocks, bonds, commodities, whatever—it’s gambling. Turning money into more money is a traditionally shadowy game, but it is a game. It’s filled with rules and conventional wisdom. Howling, chest-beating victors. Deafeningly silent and cowering losers. The dopamine hits the same. And the old gambling proverb holds true for investing: in every bet there is a fool and a thief. No middle ground. Not taking it personally is the most important lesson.
There’s a larger point to be made here: gamblers have won and lost fortunes for millennia. It’s a reflection of our very human selves: sometimes grim, sometimes glorious, always entertaining. American history is littered with gamblers masquerading as investors and inventors, businessmen and politicians, artists and athletes. Success inherently requires an unflinching understanding of risk and an insatiable desire to beat the odds. No metaphor in gambling, just the real thing.
I’m originally from South Texas. Specifically, the once-disputed area that catalyzed the Mexican-American War and the resulting Mexican Cession, the last acquisition of the unforgiving freight train America calls Manifest Destiny. The very same area that eventually supplied Lyndon Johnson—a freight train in his own right—with enough votes to steal his first Senate election. Simultaneously dusty and humid. An hour away from the beach and 20 minutes from the Rio Grande. Tacos best bought in gas stations and food trucks. Sometimes blatantly corrupt, but usually quiet. An acquired taste, the type of place you love only after it molds you.
I went to college 2,000-plus miles, one time zone, and a world away at Dartmouth. Making that move, consenting to a baptism by fire, seared into my mind that I would never know what I was capable of without getting burned. “Buy the ticket, take the ride… and if it occasionally gets a little heavier than what you had in mind, well, maybe chalk it up to forced consciousness expansion: tune in, freak out, get beaten.”
I left those New Hampshire mountains with degrees in economics and history. In true liberal arts form, my biggest takeaway was learning to learn. That and a B+ citation for proposing what my professor later named the “Mendoza Tax,” a one-point fine for every idiotic question asked during an economics exam, conceived of and communicated in a fit of annoyance at the tail-end of a three-hour final and a 27-hour cram session. Funny how life comes together in those uncomfortable moments.
More than an eerily accurate representation of how I go through the world, the “Mendoza Tax” parallels how I see the business of data science: find value-add applications for theoretical concepts. And armed with the narrative-building and argumentative rigor from my history degree, I always look to effectively communicate and contextualize the implications of results with visuals and the written and spoken word.
Not being able to bring that liberal arts approach to my work was a big source of tension for me at my first couple of jobs. I had an endless supply of interesting questions about the data landscape, but no bandwidth to get heavy into answering them, nor the technical skills to see it all the way through. That awareness drove me to data science—and to Springboard. Within three days of discovering the curriculum, I applied for the program.
Economics taught me to look for value anywhere it can be found and there’s no denying the value of the Data Science Career Track. The career coaches were invaluable in curating my resume to better reflect my experience and ambitions. My mentor Danny Wells was an incredible sounding board while I was learning Python, during my capstone ideations, and throughout their execution. The price didn’t break the bank and the job guarantee helped me sleep at night. The mock interviews forced me to prepare my pitch and served as a great recap of everything I’d learned up to that point. And I can’t possibly overstate the benefit of doing it all remotely while maintaining a job.
That last bit turned out to be less of an issue, though: time ran out on my second job, a contract I signed as a short-term solution while looking for opportunities like Springboard. Choices had to be made. Should I take another contract? Should I settle for a non-data science position? Could I reach for a junior position based on the work I’d done up to that point? If I saw this through to the end, what would I need to do to make it work?
I made it work. If you know what you’re about, these decisions make themselves. I felt I was experiencing something big—it was certainly the most difficult position in my 25 years. Success would launch me to a new and elevated plane; failure would teach me very hard and ruthless lessons about myself. The Grand Canyon was made with pressure and time. Buy the ticket, take the ride.
At that point, I shifted into high gear and grinded out a series of tasks that I only remember as a blur. The concept of “weekdays” faded into a quaint memory and all I had were hours. Hours to do this one thing. Hours to do the next. And the next. I refined my LinkedIn profile, browsing the scene with my morning quart of coffee. I made a personal website to house articles I wrote on my projects and the world contributing to them. I spent unknown time obsessing over creative application of machine learning in service of those projects. I thought deeply about the path I wanted to go down, how the narrative of these projects—and my trajectory—could change if I could piece everything together.
When you get to a certain point in your unemployment, it’s important to manage the flood of anxieties and insecurities and get them down to a manageable trickle. From there, your world becomes malleable. “Empty your mind. Be formless, shapeless, like water… you put water into a cup, it becomes the cup. You put water into a bottle, it becomes the bottle. You put it in a teapot, it becomes the teapot… water can flow, or it can crash. Be water, my friend.”
I redirected that flood of thoughts, like my addictive personality, to form a destructive crash that carried me to my goal—battered, shell-shocked, and triumphant. Truthfully, I think I’ve chosen to forget the details of those most difficult moments, buried them. “Where you been is good and gone / All you keep is the getting there.” All I remember, though I remember it well, is the feeling of the fall and the rise, the weightless drop and the g-force on the way back up.
I remained very selective throughout my job hunt. In three months of active searching, I applied to four companies, heard back from three, and got offers from two. In two days, I went from unprecedented ambiguity to a legitimate choice. I arrived at the proverbial two roads diverged in a wood. The path chose itself.
I’m not exactly recommending this approach to anybody—simply saying it worked it for me. High risk-tolerance and all of that. I took great care and consideration to formulate a pitch to each company about my ability, my potential, and how I could best contribute to their larger goals. I waited patiently until I finished the course and had more project experience before I applied. It was a targeted and comprehensive campaign. No half measures. It’s safer that way.
I’m currently a data scientist at Hypergiant, an AI consulting startup that integrates bespoke applications in the everyday processes of companies from any industry. I was initially drawn to data science as a profession for its ability to cut through the entire corporate landscape and now my reach extends anywhere there’s opportunity. I work with an impressive team of developers, designers, strategists, and other data scientists. With the help of our outstanding project managers, these diverse units come together to create some truly incredible work. I’m still impressed by their ambition to apply this formidable and evolving technology in real-world situations. I couldn’t have landed at a better spot.
They really vibed with my second capstone. I pulled lyrics for 11 artists from four genres and ran 100 different iterations of a model that predicted artists based on their word profiles. But rather than focus on what the model got right, I paid more attention to where it was wrong. Why would a country artist be confused with a rap artist? The long and the short of it, at least to me: urban decay and rural decay result in the same desolation. Perhaps we all have more in common than we’d like to think. Much like my college degrees and my professional life, I didn’t start with that direction in mind. They took the form of a cup, so they became a cup.
There’s no one way to go through this world. There’s no one kind of data scientist. Varied experiences and idiosyncrasies compel us to approach things from our own angle. No room for dogmatism in a civilization that rewards flexibility.
I hadn’t thought about my Springboard experience, and the context surrounding it, in a while. I’ve been busy working, scheming, living. I’m glad I got the opportunity to write this. It provides some closure before I let go and start “the Climb" again. A new one, technically, but really the same as it ever was.
