Skip to content

Systems

Personal methodologies, workflows, and tools.

System

Data Governance as Code: Compliance in Your PR Workflow

Governance enforced through forms and committee meetings catches violations after deployment and gets bypassed. Encoding rules as policy checks in CI/CD catches them before merge,…

System

Dashboard Design as Information Architecture

Dashboards fail when built to display data rather than support a decision. A framework built on Tufte's data-ink ratio, a layered audience hierarchy, and the…

System

Pandas vs Polars: Decision Framework, Not Benchmark

Pick Pandas or Polars by dataset scale, deployment context, and team composition, and migrate only when Pandas fails a concrete constraint, not because a benchmark…

System

When Not to Use Microservices: A Rubric

Microservices solve large-organization problems. A rubric covering team size, deployment cadence, scaling, operational maturity, and data coupling shows when a modular monolith is the better…

System

The Serverless Tradeoff Matrix

A structured framework for deciding when serverless fits: evaluate cold start latency, execution duration, invocation patterns, lock-in, observability, and team maturity rather than adopting functions…

System

Architecture Decision Records as Institutional Memory

Architecture Decision Records are short documents in version control that capture why a choice was made. The template, the pull request workflow that keeps them…

System

Tool Use Patterns in Production Agents: Taxonomy

A taxonomy of production tool use patterns for AI agents, from direct invocation to conversational negotiation, with retry, circuit breaker, timeout, and adapter strategies drawn…

System

Guardian Agent Pattern: Safety as Architecture

A dedicated guardian agent sits in the execution pipeline and approves, modifies, or blocks every proposed action against a versioned policy set of hard constraints…

System

Copilot-Agent Spectrum: AI Autonomy Framework

AI autonomy runs from passive suggestion to full autonomy. Score each task on reversibility, error cost, and frequency, then define the conditions that force the…

System

When to Fine-Tune vs RAG: Decision Framework

A decision framework for choosing fine-tuning, RAG, prompt engineering, or a hybrid based on knowledge volatility, domain specificity, latency, cost, and data volume. Start simple…

1 … 3 4 5 6