Skip to content

Forty years ago this was the whole computer — one screen, one keyboard, a hundred and twenty-eight kilobytes. Everything since has been the same idea taken apart and rebuilt smaller. Which is, more or less, what I do all day.

A 1984 Macintosh 128K coming apart and rebuilding itself as a MacBook Pro
Portfolio

I build the whole thing.

Science, markets, web, and a knowledge engine. Designed, written, and shipped by one person. If you have something hard to build, I'd like to hear about it.

01 / SCIENCE

Astrometry Pipeline

A Python pipeline that measures how bright a supernova is from telescope images. It calibrates the camera against known reference stars, then reports the supernova's brightness in standard astronomical magnitudes, with uncertainties.

4stages
g, rfilters
σpropagated error
A noisy raw telescope plate resolving into a sharp star field, with detection reticles drawn onto the brighter stars
02 / MARKETS

Quantitative Validation Framework

Combinatorial purged cross-validation with a deflated-Sharpe correction, built to catch strategies that only look profitable by luck. Nineteen strategies tested on seven years of real Databento data.

19strategies
0deployed
45folds each
A smooth glass curve shattering into hundreds of suspended shards
03 / WEB

abegerstein.com

The site you are looking at. No template and no page-builder — every component, animation, and performance decision designed and written deliberately.

7scrubbed clips
245frames each
1person
A gray wireframe filling in to become a finished rendered web page
04 / KNOWLEDGE

QuantVault Second Brain

A 10,743-entry Obsidian knowledge base spanning eight domains. Not a notes folder — a structured, cross-linked knowledge graph with a consistent authoring methodology and a retrieval layer wired directly into Claude.

10,743entries
8domains
Scattered note cards slowing and settling into an ordered three-dimensional lattice
Services

Open for freelance work.

Here's what I build most often. The list isn't exhaustive, so if you have a real problem worth solving, reach out.

S01

Custom Websites

I build marketing sites, portfolios, landing pages, and full product sites in Next.js, React, and Tailwind, then host them on Vercel or wherever you already are.

S02

Data Pipelines

I write Python that takes messy, real-world inputs and turns them into clean, structured outputs you can actually trust.

S03

Full-Stack Apps

I take an app from idea to deployed product, owning the frontend, the backend, and the database from end to end.

S04

One-Off Scripts

If there's a task you keep doing by hand every week, I'll automate it so it quietly runs itself.

S05

Anything I can build

If it's software and you have a problem worth solving, ask, because the unconventional projects are usually the most interesting ones.

About

Curious, multi-disciplinary, and only just getting started.

I'm eighteen, a recent graduate of Albuquerque Academy. This fall I head to the University of Michigan in Ann Arbor to study Aerospace Engineering with a minor in Business.

But the throughline of everything I do is software. It's the discipline I keep coming back to, because it lets me chase whatever I'm curious about: rigorous validation infrastructure for quantitative markets, a 10,743-entry knowledge architecture, scientific data pipelines that turn raw telescope frames into brightness measurements, and the portfolio you're reading right now.

I understand every line I ship. Working software is the baseline. Software you can stand behind is the standard.

18years old
4domains
ABQbased
More about me
A rough aluminum billet being machined down into a finished precision component
Approach

The principles behind everything I build.

Ship end-to-endI take projects from raw idea to live deployment: design, architecture, implementation, debugging, ops, and iteration, all of it owned. Nothing gets handed off to someone else, and nothing gets left half-finished.
Real data onlyNo synthetic series, no simulated demos. Every system is tested against actual inputs: real telescope frames, real market bars, real user flows. Edge cases are where software lives or dies.
Validate before you trustBacktests lie and mocks deceive. I build validation infrastructure alongside the system, then watch how it behaves under real load before claiming it works. Believing a false positive always costs more than verifying.
Own what shipsI understand every line of code I deliver. I make the architectural decisions, debug the failures, and can defend every design choice. If it ships under my name, I can explain exactly how it works.
A MacBook Pro closing
Let's talk

Have something worth building?

I'm open for freelance work: websites, data pipelines, full-stack apps, and custom scripts. If it's software and you have a real problem, let's talk.