Geometry of Preferences
How observed choices can be represented as latent structures—making similarities, differences and collective patterns available for computational study.
01 Research
Studying decisions as structured phenomena.
My research asks how preferences and behavior can be represented computationally—and what those representations reveal about individuals, groups and intelligent systems.
Research field
From observed
choices to latent form.
Observed → latentIndividual → collective
Behavior
Representation
Collective systems
Research interests
A connected
methodological agenda.
- 01Computational Social Science
- 02Human & AI Decision-Making
- 03Network Science & Complex Systems
- 04Representation Learning & Latent Geometry
- 05Economic Complexity
- 06Behavioral Modeling
Current themes
Human–AI Decisions
Studying the relationship between human preferences and decisions generated by artificial intelligence systems, without presuming equivalence between them.
Complex Systems & Networks
Using computational social science, network science and machine learning to investigate collective behavior and complex economic phenomena.
Research exchange
Ideas move through
institutions and disciplines.
Short-term research visits are listed separately from current appointments to keep affiliations precise.
Short-term Visiting Fellow
Knowledge Lab, University of Chicago
Short-term Visiting Fellow
Santa Fe Institute
Selected research record
Work in
development.
Statuses below follow the master CV. Works in preparation are not presented as peer-reviewed publications.
Unveiling Economic Dark Matter: Detecting and Characterizing Informal and Illicit Economic Activities through Network Representation Learning
Manuscript in preparation · targeting submission in 2026
Learning latent ideological geometry from large-scale pairwise decisions in complex social systems
Manuscript in preparation · preprint available upon request
USPTO Application No. 19/402,319
Filed November 27, 2025 · Application status, not a granted patent.
Research dialogue
Methods become stronger through exchange.
For collaboration, manuscript discussion or access to a preprint, get in touch directly.