AI systems · Geopolitical intelligence · Computational law · Knowledge graphs · OSINT/GEOINT
I build tools for reasoning through complex systems.
My background crosses several worlds: machining and automation, software engineering, linguistics, literature, Russian, musicology, military reconnaissance, scientific publishing, computational law, AI, and geopolitical intelligence.
I started with physical systems — machines, precision, automation — before moving into software and information systems. That early foundation still shapes how I work: I like things that are structured, observable, testable, and useful in the real world.
Today, my work focuses on AI-driven geopolitical intelligence and risk forecasting: combining Large Language Models, Retrieval-Augmented Generation, knowledge graphs, OSINT, GEOINT, narrative analysis, and structured event modeling to support better decisions under uncertainty.
Outside the usual software stack, I am interested in drones, SDR, sensors, maps, intelligence workflows, philosophy, language, and the deeper question of how humans build models of reality.
- AI-driven geopolitical risk forecasting
- Knowledge graphs for events, actors, locations, narratives, and risk chains
- LLM/RAG systems with source traceability and analyst workflows
- OSINT, GEOINT, drones, SDR, and sensor-informed intelligence
- Computational law and structured reasoning over legal and contractual texts
- Human reasoning, philosophy, language, and model-building
I keep coming back to the same problem in different domains:
How do we turn messy information into structured reasoning?
Contracts, reports, news, intelligence, clauses, events, maps, sensor data, and narratives are all forms of partially structured reality. I am interested in systems that make them clearer, more connected, and more useful for decision-making.
| Domain | What interests me |
|---|---|
| AI | LLMs, RAG, agents, structured reasoning, source-aware analysis |
| Intelligence | OSINT, GEOINT, HUMINT-inspired workflows, confidence scoring |
| Geopolitics | Risk forecasting, scenarios, event patterns, narratives |
| Law | Computable contracts, legal knowledge representation, reasoning engines |
| Systems | Automation, observability, infrastructure, scalable platforms |
| Sensors | Drones, SDR, geospatial signals, remote sensing |
| Humanities | Linguistics, philosophy, literature, interpretation, meaning |
I am interested in knowledge graphs not as static databases, but as reasoning surfaces: ways to connect events, actors, locations, documents, claims, sources, uncertainty, and decisions.
A large part of my work has focused on making legal documents computable: turning clauses, conditions, definitions, exclusions, limits, and scenarios into executable or inspectable systems.
My current research explores how AI can help analysts detect patterns, compare narratives, assess source reliability, and generate risk scenarios with traceable evidence.
I do not think AI replaces judgment. I am more interested in systems that make reasoning explicit, auditable, and scalable.
- How to represent complex reasoning in small, inspectable models
- How to connect LLMs with knowledge graphs and structured data
- How to detect narrative bias across sources
- How to reason over geopolitical events in space and time
- How drones, SDR, satellites, and sensors can augment intelligence workflows
- How philosophy, language, and computation shape our models of reality
From machining and automation to language, software, military reconnaissance, computational law, AI systems, and geopolitical risk forecasting.
- LinkedIn: linkedin.com/in/samuel-pouyt