03 / Insights

Implementation, verification, and the reasoning behind them.

Evidence of Capability: connect implementation and verification to design choices, trade-offs, limitations, Work and Services.

Featured Insight

EngineeringFeatured Insight

Formal Verification in the Development Loop: What Rust, Kani and Tamarin Check

Tie each verification claim to a model, implementation and explicit assumptions throughout development.

Evidence of Capability

What was checked. Why this design. What remains outside the claim.

Read the article and related work →

Eight Perspectives

Begin with the work and its design decisions.

Explore eight perspectives across formal verification, operations, legacy integration, security and AI. Each connects our five practice areas to design reasoning and the limits of its application.

Implementation, verification, and the reasoning behind them.

Formal Verification in the Development Loop: What Rust, Kani and Tamarin Check

Begin with the work and its design decisions.

Formal Verification in the Development Loop: What Rust, Kani and Tamarin Check

From Monitoring to Observability: Changing the Operating Model

Modernization Without a Rewrite: Boundaries Around Legacy APIs

Zero Trust Beyond Remote Access: Identity, Internal Traffic and Recovery

What S3 Object Lock Protects: Retention, Isolated Authority and Restoration

Six Designs for Production AI Agents: Authority, State, Execution, Observation, Evaluation and Recovery

From Tacit Knowledge to AI Workflows: Observe Decisions and Exceptions

Seven Boundaries Between AI PoC and Production: Data Through Operational Ownership

Research & Experiments

Evaluating Incus as a VMware Alternative: Start with Migration Conditions

RAG Evaluation Beyond Answers: Retrieval, Permissions, Freshness and Abstention

Architecture Notes

Six Designs for Production AI Agents: Authority, State, Execution, Observation, Evaluation and Recovery

Designing Private AI Boundaries: Data Sovereignty and GPU Operations

Modernization Without a Rewrite: Boundaries Around Legacy APIs

Zero Trust Beyond Remote Access: Identity, Internal Traffic and Recovery

Beyond Prevention: Detect, Contain and Recover from Ransomware

From Monitoring to Observability: Changing the Operating Model

Before Adding AI: Prepare APIs, Data, Authority and Auditability

Selecting AI Workflows: Frequency, Judgment, Information and Feedback

From Tacit Knowledge to AI Workflows: Observe Decisions and Exceptions

Seven Boundaries Between AI PoC and Production: Data Through Operational Ownership

Engineering

Formal Verification in the Development Loop: What Rust, Kani and Tamarin Check

What S3 Object Lock Protects: Retention, Isolated Authority and Restoration

Long-Term Metrics Beyond Prometheus: Thanos Retention, HA and Queries

Vector and OpenTelemetry Collector: Drawing Collection Boundaries

Managing Kubernetes Updates with Nix: Version Pinning and Recovery Boundaries

From VLANs to Enforcement: Inspecting East-West Traffic

From COM to an F# Domain API: Options, Results and State at the Boundary

Budgeting Local LLM VRAM: Beyond Model Weights

Field Notes

When Kani Exploration Stalls: Record the Boundary Before Narrowing It

The Gap Between Tamarin and Implementation: Recording Correspondence

After OpenTelemetry: Can the Signals Explain a Failure?

Who Notices Monitoring Failure? Observation Outside Kubernetes

Cloudflare Zero Trust and Entra: Assigning Policy Decisions

Before Setting Object Lock Retention: Recovery Windows, Cost and Deletion

Letting Agents Operate SaaS: Whose Authority Commits the Action?

Have a technical problem worth writing about?