Hi, I’m Max. I’m a high school senior from the southern tip of Texas, looking to major in applied math and do quantitative research. I love learning for its own sake and follow my curiosity wherever it leads me, be it ramification theory or Sumerian literature.
Right now I’m part of a team at UTRGV that resolves a 2013 conjecture about determinants, building Heatwave (an app that finds the most tolerable walking route instead of the shortest), and leading a NASA HUNCH team that predicts failures in lunar rovers.
My school ran out of math classes to put me in after Calculus I, so since then I’ve learned math through research, MathILy, and self-study.
Email me at mmoylemaish@icloud.com, or find me on GitHub and LinkedIn. My CV has everything in one place.
Sunlit minutes, 400 random routes
Median across all 18,564 routes: 8 min.
Fig. 1. Shade is free on a grid. When you can only walk along streets, distance is taxicab distance, and the shortest route is rarely unique. There are 18,564 routes from A to B, and every one takes 18 minutes. They aren’t equally sunny, though. The gray lines are 400 of these routes chosen at random; a typical one keeps you in the sun for 8 minutes. The red line , found with dynamic programming, is the shadiest of all of them, resulting in only 2 sunlit minutes.
Each stretch of street between two corners takes a minute to walk. Stretches lined with trees are shaded; the rest are in the sun. Drag A or B, or .
A 2013 conjecture says that a family of integer matrices built from Eisenstein series always has determinants that are powers of three, up to sign. I’m part of a team at UTRGV working on it and on the structure behind it, and we’re aiming to post a preprint by the end of the semester. I’ve worked on proofs and written the Python code we use to compute and check the matrices.
The research page has more detail.
MathILy is a five-week summer program in proof-based math at Bryn Mawr College, and it admits about 7% of applicants. We worked on combinatorics, algebra, polytopes, and more. We were given statements and setups and had to find the proofs ourselves, sometimes spending more than a week on one problem.
Heatwave gives walking directions that avoid the heat. For each street, it estimates how hot the walk will be from the current weather, the height of the sun, the shade from trees, and whether the ground is asphalt, concrete, or grass. Then it finds the fastest route, the coolest route, and a balance between the two. I built it on my own with Python, GeoPandas, OSMnx, rioxarray, and Node.js, and people have used it to plan more than 10,000 walks across the United States.
You can try it at heatwaves.app or read how it works.
I led a team of three that built an RFID system for checking supplies in and out on long deep-space missions, designed for the logistics module of NASA’s Gateway station. I wrote the React and TypeScript app, the Node.js server, and the PostgreSQL database. This year I’m leading a new team that’s working on predicting failures in lunar rovers, and I designed the system, its 17 data schemas, and its dashboard.
In the summer of 2026, I spent a week at the Summer Academy for All: Machine Learning Edition, run by UT Austin’s Learning Directed Operating Systems lab. I built a small classifier that predicts how slow a computer’s input and output requests will be, and it had to make each prediction in under 100 nanoseconds. While working with the lab’s data, I found a leak: the model could see information it shouldn’t have, which made the accuracy in the lab’s notebook look better than it really was. The researchers hadn’t caught it, and they adopted my fix.
One way doctors measure someone’s risk of falling is to stand them on a platform that shifts under their feet and record how they keep their balance. The machines that do this are expensive, so for our engineering capstone, my team of four is building a cheaper, portable one. I lead the research and the math behind it, and we plan to have a working prototype by April 2027.
In my engineering practicum, I led student teams on projects for real clients. For NerdVana, a local education startup, we built a prototype of an augmented-reality sand table, and I worked out how its sand should drain. The next year, we built the framework and homepage for NerdVana’s new website in a month, using Astro and Keystatic, and wrote its documentation. I also led a team that maintained the City of Progreso’s website and drafted a redesign, and I built a website for a friend who tutors physics.
I’m the vice president of my school’s chapter this year, after a year as treasurer. I design our year-long curriculum of 16 meetings, give the lectures, and write the problems for the three competitions we run each year: a Differentiation Bee, an Integration Bee, and a general math contest. Membership roughly tripled in a year, and we raise our own money for competition trips.
I founded and ran a study group of more than 40 students, with our teachers’ approval, mostly for AP World History but also for calculus, physics, and computer science. These days I also tutor middle schoolers in math every week.
I play the piano (badly), run, and listen to a lot of classical and electronic music. I like art, ancient Near Eastern history, and Sumerian literature, and there are two Sumerian things hidden on this site. I keep half an eye on golf and tennis.