AI Product Engineer

Erik Mejia

I build AI products where software, intelligence, and interface design need to work together.

Since 2011 I've built software across education, government, and commerce. In 2017 I donated a nationwide weather-alert platform to the Dominican Republic that served 50,000 active users in one week during hurricanes Irma and Maria, later surpassing 100,000 installs without a single outage.

Today I focus on AI product engineering: shaping product direction, building the systems behind it, and refining the experience so the result is useful, understandable, and worth shipping.

Active projects

What I'm building now, with a focus on AI products where software, intelligence, and interface design need to work together.

Flowcost

A planning tool for modeling AI workflow cost, latency, and infrastructure tradeoffs before you build.

Didastik

In development

An AI learning platform that turns study material into guided review, quizzes, and notes people can actually use.

Selected work

A broader sample of work I led end to end across design and engineering, from web products and Android apps to operational tools.

Graphite

Live product

A Webflow app for auditing sites, improving SEO structure, and fixing optimization work at scale.

2024

Epidemia RD

Legacy preview

A public web app for tracking COVID-19 data in the Dominican Republic during the pandemic.

2020

Rastreado

Legacy preview

A logistics platform for modernizing delivery operations with GPS tracking, digital signatures, and operational visibility.

2018

Onamet

Open-source repo

A nationwide weather-alert platform for the Dominican Republic used heavily during hurricanes Irma and Maria.

2017