↓Skip to main content

Projects

These are some of the projects I’ve truly enjoyed working on! You can use the tags and categories sections to find something specific.

Seven 8-bit ponies racing mid-race past a cheering crowd in Super Pony Picker

Super Pony Picker: Random Name Picker

·5 mins
A free and fair random name picker disguised as an 8-bit horse race. Enter up to 20 names, choose whether to pick the winner or the last one (1, 2 or 3 people) and watch the ponies race through turbos, stars, mud and dice rolls to the podium. Built with p5.js, the Web Crypto API and live chiptune sound, it runs entirely in your browser.
Exhibition view of Visiones inestables: a projected AI-generated image of the museum, a projected quote and a lit photograph in a dark gallery

Visiones inestables, by Mónica Páez

·3 mins
Visiones inestables (Unstable Visions) is an artwork by Mónica Páez that uses AI to imagine unstable futures of the Art Museum of the Universidad Nacional de Colombia, from photographs of the museum and texts about the institution. Exhibited at the museum itself in June 2026. I had the privilege of collaborating on it with Juan David García Arteaga.
The empty exhibition space, with white walls, large windows and a ceiling projector, that La traición necesaria describes and recreates

La traición necesaria, by Mónica Páez

·3 mins
La traición necesaria (The Necessary Betrayal) is an artwork by Mónica Páez inspired by the adage traduttore, traditore: an AI describes the exhibition space and creates a new image from that description, again and again. Exhibited in March 2026 in an exhibition hall at the Universidad de los Andes. I had the privilege of collaborating on it with Juan David García Arteaga.
Cartoon robot holding a wrench and giving thumbs up

E-Commerce Agent with Ollama & LangChain

·64 mins
A conversational agent for e-commerce built with Ollama, LangChain, and ChromaDB, comparing three open-source LLMs — Qwen3 4B, Llama 3.2 3B, and Phi-4-mini — for product search, order tracking, recommendations, and RAG-powered FAQ. Benchmarks tool-calling accuracy, multi-intent handling, and response time across all three models.
Illustration of the Prisoner's Dilemma on a network with a cooperative green agent and a defecting red agent

Prisoner’s Dilemma in Dynamic Networks

··43 mins
Agent-based simulation of the iterated Prisoner’s Dilemma on adaptive networks. Explores how cooperation emerges through imitation (Fermi Rule), dynamic rewiring, and Pareto wealth across Erdős-Rényi and Barabási-Albert topologies.