Carlos Daniel Jiménez · AI engineering & applied research
Music, meaning,
and reliable AI.
I help engineering teams understand where their AI systems fail, evaluate the evidence behind their answers, and decide what to change.
My research explores music through embeddings and narrative arcs: how bands evolve, how albums develop ideas, and what a computational interpretation preserves or misses. Those questions also inform my work on agent memory, retrieval, and evaluation.
Music, embeddings, and narrative arcs
How does music develop an idea?
I explore how a band's lyrics change and how songs relate through close reading, embeddings, and narrative-arc analysis. A recurring theme can change its meaning; an album's ending can reshape how we hear its beginning.
The essays examine the Beatles' lyrical evolution, multilingual lyrics in Aquamosh, and what an agent needs to remember across an album. Each experiment describes the evidence it uses and the limits of its interpretation.
AI Software Engineering
Narrative Arcs in Music
For teams building with AI
Your system has answers. Which ones can you trust?
A focused diagnostic examines one RAG or agent workflow, typically through 20–30 cases over two calendar weeks. You receive an error analysis, an evidence review, resource accounting where traces support it, and priorities for the next engineering decision.
Architecture reviews, inference optimization, and ongoing technical advisory can follow when the findings call for them.
See the scope and request a conversation →