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.

165

Articles

Likes

383

Views

Comments

Latest from Data Bricklaying Diary

100 recent articles max

Data Bricklaying Diary

Oct 13, 2026 · Industry Insights

Don't Wait for Perfect Data: Phased Gates for Ontology Pilots

The article argues against waiting for full data governance before starting ontology pilots, proposing a four-phase gated approach—concept verification, sample verification, continuous query, and business execution—with explicit three-state data quality handling (satisfied, unsatisfied, unknown) and clear separation of query conclusions from execution permissions, using a warehouse shortage assistant case study.

TOCTOUW3C DQVdata governance

0 likes · 16 min read

Don't Wait for Perfect Data: Phased Gates for Ontology Pilots

Data Bricklaying Diary

Oct 12, 2026 · Artificial Intelligence

Agent Observability: Correlating Model, Tools, and Business Receipts to Verify Task Completion

This article argues that Agent observability must go beyond traditional microservice tracing by correlating model calls, retrieval, tool invocations, and control-plane events to task-level identifiers, tracking phases, waits, costs, evidence, and versions, and verifying outcomes with reliable business receipts rather than relying solely on API success metrics.

AI EngineeringAgent ObservabilityOpenTelemetry

0 likes · 14 min read

Agent Observability: Correlating Model, Tools, and Business Receipts to Verify Task Completion

Data Bricklaying Diary

Oct 12, 2026 · R&D Management

Start with One Scenario or Build Full Ontology? Preventing Pilot Silos

The article advocates starting enterprise ontology with a verifiable business capability like supplier qualification checks, defining object identity, scope, shared definitions, and maintenance responsibilities upfront, then using a second scenario to test reuse boundaries while tracking dependencies for future changes.

architecture decision recordsdata governancedependency tracking

0 likes · 14 min read

Start with One Scenario or Build Full Ontology? Preventing Pilot Silos

Data Bricklaying Diary

Oct 11, 2026 · R&D Management

Who Owns Enterprise AI? Six Responsibility Types for Real Business Value

This article argues that enterprise AI requires six distinct responsibility types—strategic, business value, employee change, platform, risk, and operations—each with a single accountable owner, and provides a decision matrix with escalation paths to replace vague RACI charts and ensure AI delivers sustained business value.

AI governanceCross-Functional TeamsDecision Matrix

0 likes · 25 min read

Who Owns Enterprise AI? Six Responsibility Types for Real Business Value

Data Bricklaying Diary

Oct 10, 2026 · Artificial Intelligence

Synthetic Data Isn't Free Real Data: When to Use & How to Validate Without Bias

This article explains that synthetic data cannot replace real evidence, detailing when it's appropriate (rare scenarios, privacy, simulation), five validation dimensions (authenticity, validity, diversity, privacy, task gain), strict separation between training and evaluation sets, auditability requirements, and the need for reproducible task gains on real independent benchmarks before adoption.

auditabilitydata generationdata validation

0 likes · 28 min read

Synthetic Data Isn't Free Real Data: When to Use & How to Validate Without Bias

Data Bricklaying Diary

Oct 9, 2026 · Artificial Intelligence

Why AI Agents Need Conclusion Gates: Verifying Business Decisions Beyond Plausible Outputs

This article argues that AI agent outputs resembling conclusions require systematic verification gates in business systems, detailing how to separate facts from inferences, validate evidence scope and freshness, enforce access controls, and distinguish conclusion release from action authorization—illustrated with procurement examples and adversarial test cases.

AI agentsaccess controladversarial testing

0 likes · 17 min read

Why AI Agents Need Conclusion Gates: Verifying Business Decisions Beyond Plausible Outputs

Data Bricklaying Diary

Oct 9, 2026 · Artificial Intelligence

Beyond Tool Calls: FDE's Responsibility Chain for Production AI Agents

This article details FDE's framework for integrating AI agents into enterprise workflows, covering identity delegation, data egress inventories, action control with six record types, evidence binding across trace and audit layers, failure handling with four exception-path questions, compliance checkpoints embedded in state transitions, a five-party responsibility model, and ten required security deliverables.

AI agent governanceaction contractsaudit trails

0 likes · 26 min read

Beyond Tool Calls: FDE's Responsibility Chain for Production AI Agents

Data Bricklaying Diary

Oct 8, 2026 · Databases

Master Data Aligned but Reports Wrong? The Three-Layer Fix: Identity, History, Query Logic

The article explains why unified equipment codes still cause report errors due to missing temporal equipment-workstation relationships, and outlines a three-layer diagnostic (identity alignment, historical relationships, query semantics) advocating reuse of existing master data and history tables, complemented by ontology for shared semantics, with regression testing to verify fixes.

SCD Type 2data governancemaster data

0 likes · 14 min read

Master Data Aligned but Reports Wrong? The Three-Layer Fix: Identity, History, Query Logic

Data Bricklaying Diary

Oct 7, 2026 · Backend Development

Agent Error Recovery: Why Retries Alone Fail — Idempotency & Reconciliation

This article explains why automatic retries after agent timeouts can duplicate business actions, and details how to distinguish call failures from executed side effects, implement execution-side idempotency guarantees, reconcile using original action identifiers, and define safe recovery rules with human escalation when outcomes remain unknown.

Agent SystemsIdempotencydistributed systems

0 likes · 17 min read

Agent Error Recovery: Why Retries Alone Fail — Idempotency & Reconciliation

Data Bricklaying Diary

Oct 7, 2026 · R&D Management

Mission Engineering 12: Scaling from One Cup to Ten — What Carries Over to Complex Delivery?

This article explores how mission engineering methods scale from brewing a single cup of coffee to serving ten people and fulfilling complex orders, emphasizing that while definitions and recipes can be reused, goals, timing, resource capacity, handoffs, and success evidence must be re-validated for each new scenario, using OPM and ontology to trace cross-system collaboration and emergence.

Cross-team CollaborationEmergenceMission Engineering

0 likes · 24 min read

Mission Engineering 12: Scaling from One Cup to Ten — What Carries Over to Complex Delivery?