Selected Work

Applied work across research, engineering, and teaching.

I'm a PhD engineer who has done forecasting for Peru's Central Bank, taught computer science at the university level, published ML research, and built a set of Rust tools I actually use in my own work. This page is an honest record of what I've shipped.

What I Do

Three areas where I have real depth.

Applied Machine Learning

I've used ML in production settings — not just tutorials. My Central Bank work involved building nowcasting models for GDP and macroeconomic indicators using state-of-the-art forecasting methods.

  • • Forecasting and nowcasting
  • • Bayesian variable selection (MCMC)
  • • PyTorch, Keras, scikit-learn
  • • Research-to-deployment pipelines

Developer Tooling in Rust

I build CLI tools and infrastructure for working seriously with AI. Written in Rust for stability. Designed from real friction, not imagined use cases.

  • • Codebase flattening (Yggdrasil)
  • • Conversation control (FUR)
  • • Schema-flexible data (Membrane)
  • • LaTeX bibliography tooling (Elaine)
See Tools

Web Engineering & Deployment

I've built and deployed full-stack web products for real operators — including a live retail brand and an independent commerce launch. React, Next.js, end-to-end.

  • • Next.js / React
  • • Payment integration
  • • Operator-owned infrastructure
  • • Rapid deployment cycles
Deployments

Things that shipped and ran.

Central Bank of Peru — ML Consulting

Delivered forecasting and nowcasting models for macroeconomic indicators including GDP. Introduced state-of-the-art ML methods to an institutional research team.

University Teaching — CS and Python

Taught data structures and algorithms in C++ at UTEC Lima, and a foundations-of-programming course in Python at UP Pacífico. Both at the undergraduate level.

Halo Effect — Commerce Deployment

Owned infrastructure supporting a live retail brand with full payment integration and operational control.

haloeffect.vercel.app →

PuntoMalla — Operator Launch

Digital commerce environment enabling an independent operator to start direct sales without platform dependency.

puntomalla.vercel.app →
Research

Work in progress, in the open.

Variable Selection via Coordinate Transformation

A novel variable selection algorithm that reframes the problem by changing coordinates — moving from the variable space into a corresponding metadata space. Submitted to Applied Soft Computing. Previously revised following peer review at Knowledge-Based Systems.

Bayesian Methods · MCMC · Variable Selection · Applied ML
Explorations

Experiments I built to find out if something was possible.

Moxi — Voxel Programming Language

A programming language designed so that AI models can generate structured 3D environments more reliably — closer to natural language than traditional 3D APIs, but precise enough for machines to reason about. Still in active design.

Work in Progress · Language Design · 3D · AI Tooling
View on GitHub

The Fractal Prince — Video Game

A recursive platformer I built as an experiment in spatial logic. Looping geometry and constrained traversal quietly reshape how a player moves through the world. Built and shipped end-to-end.

Game Design · Spatial Logic · Shipped
Play

I'm currently open to contract work, research collaborations, and the right full-time role. If something here is relevant to what you're building, I'd like to hear about it.

Get in Touch