Author

Data Bricklaying Diary

Records practices, thoughts, and pitfalls on the data grunt-work journey, sharing content on data platforms, data analysis, data processing, data governance, knowledge graphs, and more. Less theory, more hands‑on, making complex data technologies simple.

Articles

Likes

Views

Comments

Latest from Data Bricklaying Diary

86 recent articles

Data Bricklaying Diary

Aug 27, 2026 · R&D Management

FDE Series: Why Enterprise AI Needs a Problem-to-Production Responsibility Loop

The FDE series introduces a responsibility loop for enterprise AI deployment, connecting semantics, agents, business research, engineering, production validation, handover, and product feedback, arguing that successful AI requires an engineering responsibility model embodied by Forward Deployed Engineers.

AI DeploymentAI ProductionEngineering Responsibility

0 likes · 14 min read

FDE Series: Why Enterprise AI Needs a Problem-to-Production Responsibility Loop

Data Bricklaying Diary

Aug 26, 2026 · Artificial Intelligence

AI Evaluation Sets ≠ Training Holdouts: Golden Data, Hard Cases, Contamination & Regression

This article explains why enterprise AI evaluation requires purpose-built datasets — golden baselines, challenge cases, red-team tests, regression suites, and online replays — designed from capability questions, risk lists, and real failures, with structured answers, contamination governance, tiered LLM judging, and version-locked regression pipelines gated by risk-based release thresholds.

AI evaluationLLM-as-a-Judgechallenge set

0 likes · 24 min read

AI Evaluation Sets ≠ Training Holdouts: Golden Data, Hard Cases, Contamination & Regression

Data Bricklaying Diary

Aug 24, 2026 · Artificial Intelligence

Codex Harness Deconstructed: Agent Runtime Beyond the Execution Loop

This article analyzes Codex Harness's architecture, revealing it as a full Agent Runtime with Core, state model, App Server control plane, event approval, Goal, Queue, Memory, and multi-agent collaboration—far beyond a simple model-tool execution loop—and highlights gaps for enterprise adoption like business semantics and governance.

AI agentsAgent LoopAgent Runtime

0 likes · 15 min read

Codex Harness Deconstructed: Agent Runtime Beyond the Execution Loop

Data Bricklaying Diary

Aug 24, 2026 · Fundamentals

Ontology Isn't Esoteric: You Already Process Its Raw Materials Daily

This article explains that ontology modeling uses familiar business objects, attributes, relationships, and constraints from daily data and process work, but adds cross-system unified semantics, explicit computable constraints, machine verification, and continuous operations to enable data, rules, systems, and AI agents to collaborate on a shared business world.

AI readinessOWLSHACL

0 likes · 20 min read

Ontology Isn't Esoteric: You Already Process Its Raw Materials Daily

Data Bricklaying Diary

Aug 23, 2026 · Fundamentals

From Business Research to Ontology Modeling: 7-Step Framework for Verifiable Business Models

This article details a seven-step methodology to transform business research findings into verifiable business models ready for ontology engineering, using an order fulfillment risk case study to illustrate object identification, rule definition, evidence mapping, and validation through competency questions and boundary samples.

Data Governancebusiness modelingbusiness rules

0 likes · 28 min read

From Business Research to Ontology Modeling: 7-Step Framework for Verifiable Business Models

Data Bricklaying Diary

Aug 22, 2026 · R&D Management

From Vague Requirements to Verifiable Business Scenarios: A 7-Step Ontology Research Method

This article outlines a seven-step business research methodology for ontology modeling, transforming vague requirements into clear, verifiable business scenarios by distinguishing facts, rules, judgments, and hypotheses, evaluating candidate opportunities, defining value narratives and boundaries, and producing a scenario definition card as a modeling input.

Data Governancebusiness researchknowledge engineering

0 likes · 29 min read

From Vague Requirements to Verifiable Business Scenarios: A 7-Step Ontology Research Method

Data Bricklaying Diary

Aug 21, 2026 · R&D Management

Feature Complete ≠ Production Ready: Why AI Coding Demands Engineering Discipline

This article argues that AI can rapidly generate functional code but cannot lower the engineering bar for production readiness, which requires risk-matched baselines, independent verification, and accountable gates — illustrated through a batch-import example showing the gap between happy-path code and real-world constraints like retries, idempotency, observability, and rollback.

AI-assisted developmentIdempotencyengineering gates

0 likes · 21 min read

Feature Complete ≠ Production Ready: Why AI Coding Demands Engineering Discipline

Data Bricklaying Diary

Aug 21, 2026 · Fundamentals

Why Ontology Projects Must Define Competency Questions First: Verifiable Acceptance from Day One

This article explains why ontology projects should define competency questions before modeling, detailing how these verifiable business questions constrain scope, expose data gaps, guide modeling trade-offs, and enable repeatable acceptance testing through traceability to data, rules, evidence, and test cases.

Knowledge Representationcompetency questionsmodel acceptance

0 likes · 22 min read

Why Ontology Projects Must Define Competency Questions First: Verifiable Acceptance from Day One

Data Bricklaying Diary

Aug 20, 2026 · Artificial Intelligence

One Dataset Fits All? Why Training, Eval, RAG & Agent Data Must Be Separate

The article explains why AI systems need four distinct data products—training sets, evaluation sets, RAG knowledge bases, and Agent contexts—each with separate purpose, structure, timeliness, isolation, and acceptance criteria, warning that reusing a single dataset creates false quality metrics and operational risks.

AI data productsAgent ContextData Governance

0 likes · 20 min read

One Dataset Fits All? Why Training, Eval, RAG & Agent Data Must Be Separate

Data Bricklaying Diary

Aug 19, 2026 · R&D Management

AI Redraws Responsibility Boundaries: The New Full-Stack Isn't About Doing Everything Alone

The article argues that AI lowers cross-stack execution costs, prompting a shift from frontend/backend separation back to full-stack—redefined as end-to-end business capability ownership with AI assistance—while professional judgment, risk decisions, and independent verification remain essential and must be allocated based on task complexity and risk.

AI-assisted developmentbusiness capability teamsfull-stack development

0 likes · 20 min read

AI Redraws Responsibility Boundaries: The New Full-Stack Isn't About Doing Everything Alone