Pipelines · Identity · Privacy
Every cleaning rule is a claim about its consumer
Where data comes from, what is done to it on the way, who it belongs to, and what the law says about using it. Pipelines and warehouses, ontologies and retrieval, identity across systems, and the privacy rules that follow the record wherever it goes.
A record that joins to the wrong person is wrong everywhere it travels — the join is where the trust is decided.
The writing here treats data as something with an owner and a history. A pipeline encodes decisions about who will read its output, an identity graph decides whether two records are one person, and a privacy rule is only as good as the evidence that it was followed. It spans industries: clinical records, marketing lists, fraud signals and the corpora AI models read.
The frame
Four rules the data writing keeps coming back to
Read who owns the join →Cleaning is a decision
Every dedupe, merge and filter is a claim about the consumer. Write it down where the consumer can check it.
One person, one record
Marketing, billing, care and fraud all fail the same way when two records for one person never join.
Collect what was chosen
Data a person gave on purpose holds up as cookies, consent rules and regulators tighten.
Privacy is evidence
A policy says what should happen; a log, a lineage and a deletion record show what did.
Flagships
Where to start
Data Pipelines for Machine Learning
A second sensor doubled the evidence and the merge hid the thing it was watching for, because every cleaning rule is a claim about the consumer, and here the consumer was a model.
Technology Responsibility
A compliance and security reference for anyone who holds other people's data: what the main regimes require, how the obligations differ by sector and jurisdiction, and what a team has to be able to show.
Field notes
Data and privacy writing
See all →- Technology Responsibility: A Compliance and Security StudyData
- Clinicians and AI: What They Actually AskHealthcare AI Evidence
- MARA, the Medical Agentic Resident Assistant: Four Generations, One LineHealthcare AI Design
- MAX4 and MARA: The Ladder Starts at OffMAX4, Trust & Clinical AI
- MAX: The AI Resident, and the Road to MARALocal-First AI & Healthcare
- MARA: A Medical Agentic Resident Assistant That Learns From Every Encounter and Never Makes the DecisionHealthcare AI Design
- Epic Migration Planning: Twelve Thousand Questions, One DateHealthcare IT Program Planning
- Data Pipelines for Machine Learning: The Duplicate That Hid the HeartbeatData Engineering & ML Pipelines
- UML: The Boxes Were Right and the Lines Were WrongModelling Notation & Diagrams
- Retrieval-Augmented Generation: Confident, Well Cited, and WrongRetrieval & Search Architecture
- Domain-Driven Design: The Word the Code Got WrongArchitecture & Domain Modelling
- The Flat LineAI Writing & Detection
- Reasoning Without a ReasonerReasoning & Ontologies
- The Same MAX in the MorningMemory & Identity
- You Were Already in the FileAgentic AI
- Classic RAG vs GraphRAG vs Agentic RAG: Choosing the Right Retrieval PatternProtocols & Architecture
- Your AI Strategy Has a Single Point of Failure: Your DataStrategy
- Dashboard BI vs Agentic Analytics: The Next Wave of Business IntelligenceTools & Ecosystem
- Vectorless RAG vs Traditional RAG: Navigating by Structure Not SimilarityProtocols & Architecture
- Racing Toward the Future: The Accelerated Evolution of Machine LearningCloud / DevOps
- Unlocking Healthcare Data: The Intersection of HL7, FHIR, and HIPAA ComplianceHealthcare
- Wiring the Clinical RecordHealthcare
- Connecting the Record and the ContentData
- Elevate Your Data Strategy with OntologiesData
- From Data Revolution to Data Protection: Safeguarding PII and PHI in TechnologyData
- Securing the Future: Technology's Role in Data Privacy ComplianceData
- Who Owns the JoinData
- The Window That Only Opens One WayNarrative
Engage
Records that never quite join?
A look at where your data comes from, which records should be one person and are not, and what you could produce if a regulator or a customer asked what happened to theirs.
This site: HTTPS only · no third-party trackers · no analytics cookies · security.txt