I lead engineering teams and build cloud platforms, distributed services, and developer tools. My public projects span AWS infrastructure, reusable agent workflows, and experiments with local AI and vision models.

Public AWS announcements are matched against configured services and risk phrases, mapped to Slack routes, stored in DynamoDB, and delivered through SQS.

Open walkthrough page · captions and downloads · Architecture · Repository

A local development platform for building and testing distributed services with Docker Compose, Kubernetes, messaging, databases, and observability. Relay, a webhook delivery service, is its first application, with more services planned. A separate Terraform workflow supports brief AWS validation runs.

View the full-size system diagram ↗

Local development comes first. The live EKS, RDS, and MSK relay proof is pending.

Explore the platform · Relay service · Repository

Reusable workflows for Codex and Claude Code, covering project planning, architecture decisions, repository checks, and Git publishing. Each skill packages instructions and supporting tools with explicit verification and permission boundaries.

View the full-size skills map ↗

Browse the skill catalog · Skill inventory · Repository

A local dog-photo investigation prototype experimenting with local vision models and agent-directed tool use. Scout uses Qwen3 to choose investigation steps, while a separate vision model ranks visual breed matches and records evidence for the assessment.

Scout running a local vision model during a real demo investigation.

Screenshot and photo credits · Beagle photo by sannse · CC BY-SA 3.0

Completed prototype; model experiments closed. Findings document the reporting limits.

Explore the architecture · Findings and limits · Repository

• CliSpecForge: a compact Python reference implementation that turns a Markdown CLI spec into files that can be inspected before guarded writes.
• Compact Theme: a framework-free HTML, CSS, and JavaScript theme used across my GitHub Pages sites.

• Platforms and distributed services: building a shared local platform for multiple applications, with clear service boundaries, observability, and controlled AWS validation.
• Agent-assisted development: creating reusable workflows for planning, architecture decisions, repository checks, and publishing, with explicit permissions and changes checked against specifications and tests.
• Local AI and vision models: exploring model behavior through small applications, visible tool execution, and measured results that document usefulness and limitations.