
Pablo Ríos
Software Engineer
About
Engineer working on AI infrastructure and data problems.
I’ve worked on products across finance, education, and the food industry.
Now at Metal, where I work closely with private equity firms to build and deploy AI systems around their data and the way they work.
Interested in pushing what’s possible with agentic systems at scale.
If you're tackling similar challenges or just want to connect, I'd love to hear from you.
Projects
Work to Knowledge Skill
An AI agent skill for learning from everyday work, from coding to research and planning. Turn journals into debriefs, concept notes, and learning paths in Markdown, with optional Obsidian.
AIAgent SkillsLearningMarkdownSimilar Earth
Open source search engine that finds every place on Earth that looks like a given location. Built on AlphaEarth embeddings with an in-memory global grid and on-demand 10m tile refinement.
AlphaEarthEmbeddingsGoGeospatialOpen Source
Writing
There’s a pope in your database
How Raft keeps distributed systems from splitting up or getting stuck.
The show must go on (even when the network doesn’t)
How to sync thousands of phone screens without relying on a live network.
A search engine for places that look alike
How Similar Earth uses AlphaEarth embeddings, an in-memory grid, on-demand tile refinement, and aggressive caching to find every place on Earth that looks like a given location in about two seconds.
Fortune 500 AI Disclosure Analysis: How America's Largest Companies Talk About AI in SEC Filings
An analysis of AI mentions across Fortune 500 SEC 10-K filings from 2022 to 2025: who talks about AI, who frames it as risk vs. opportunity, and whether disclosure correlates with performance.