Systems That Decide What Matters
Building systems that decide what's real — and help humans decide what matters
North Star
Designing systems that remember, reason about, and act on real-world entities under uncertainty
Building Systems That Decide What’s Real
This is a long-term research and learning log focused on how modern systems decide what is real, what is reliable, and what humans (or machines) should act on. Gleaning insight from currently business intelligence reporting.
Theme: Designing systems that remember, reason about, and act on real-world entities under uncertainty.
Agents should act only on what they understand, remember, and trust.
Context
Most modern software systems are built on top of fragile assumptions about identity. People become rows in tables. Places become strings or coordinates. Companies become UUIDs. As long as humans remain in the loop, these abstractions mostly hold. But as decision-making shifts toward automation and AI systems, those shortcuts begin to fail, often in costly, invisible ways.
What I am working toward is a deeper understanding of how identity, context, and evidence come together to allow systems to safely answer basic questions like “what is this?”, “have we seen this before?”, and “can we act on this information?”.
Research Focus Areas
At the intersection of identity, representation, and decision systems
Identity & Entity Resolution
Understanding how systems determine "sameness" across fragmented, noisy, and evolving data sources. Identity is a belief, not a fact.
Data Representation & Embeddings
How we represent entities and relationships shapes what systems can learn and decide. Representation matters more than models.
AI Inference & Decision Systems
Building systems that don't just predict—they decide and act. Decisions are different from predictions.
Trust, Confidence & Provenance
Systems must track where information comes from, how confident they are, and why they believe what they believe. Confidence should be first-class data.
Human Attention & Judgment
Technology shapes what humans pay attention to and how they make decisions. Systems shape human attention. Automation must remain accountable.
Guiding Principles
Core beliefs that shape this research
Identity is a belief, not a fact
Systems must acknowledge uncertainty in entity identification. Representation matters more than models.
Confidence should be first-class data
Uncertainty must be tracked, not hidden. Decisions are different from predictions—acting requires more than probabilities.
Automation must remain accountable
Systems shape human attention and must explain their decisions. What we show influences what people decide.
Where This Matters
From business intelligence to autonomous systems
Business Intelligence
Humans making decisions based on data—systems must present truth with context, confidence, and provenance.
Autonomous Systems
Agents and robots making decisions—systems must understand, remember, and trust what they act on.
Integration Layer
MCP and A2A define how agents access tools. We provide the "reality check" tool agents call before acting.
Join the Journey
This is a 3-5 year commitment to understanding how systems decide what's real. If you're working on similar problems—entity resolution, knowledge graphs, AI trust, or decision systems—let's connect.
Follow the research: