Will Yennie

Full-stack and infrastructure work for a Boulder-based web design consultancy with 600+ clients across its portfolio.

• Worked directly with clients to scope requirements and ship features and fixes across the agency's portfolio of sites.
• Built a Python tool to automate vulnerability monitoring across 100+ client sites, flagging outdated or compromised dependencies.
• Built a reusable Google Reviews component (Google Places API) that any client site could embed.
• Hardened DigitalOcean infrastructure for 40+ client websites: security updates (Apache, MySQL), SSL provisioning, and monitoring.

stack: Python · JavaScript · PHP · DigitalOcean · MySQL

A CLI for dispatching multiple Claude coding agents in parallel, each isolated in its own git worktree with declared path-ownership lanes.

• Designed a manager / specialist / reviewer pattern with runtime enforcement that prevents parallel agents from writing outside their declared file paths.
• Implemented persistent per-specialist memory with cross-project and per-project scopes, auto-captured after each mission via structured model output.
• 600+ tests, published to PyPI, MCP server integration, and GitHub Actions support.

stack: Python · claude-agent-sdk · Typer · Pydantic · asyncio

A campaign organizer for tabletop RPG game masters: a tree of pages with a WYSIWYG editor, plus an AI chat scoped to whichever page you are on.

• Built a tree-structured page system with an inline WYSIWYG editor, so game masters nest factions, cities, NPCs, and session logs as one navigable hierarchy.
• Built a context-aware chat panel that scopes each answer to the current page, so a question about one NPC pulls from that page rather than the whole campaign.
• Streamed AI responses token-by-token over WebSockets, and ran a beta with real users that measured a 20% lift in satisfaction across the cycle.

stack: Rails · WebSockets · AI APIs · WYSIWYG editor

Built a chatbot feature for the official ZenML VSCode extension.

• Integrated OpenAI and Anthropic Claude APIs for context-aware chat scoped to the user's selected ML pipeline.
• Built the chat UI: dropdowns, interactive controls, per-conversation context selection, and state management.
• Implemented token-by-token message streaming so responses arrive live instead of as a single blocking call.

stack: TypeScript · VSCode Extension API · OpenAI API · Anthropic Claude API

Open-source observability framework for AWS Lambda. Serverless function health via metrics and traces.

• Implemented W3C TraceContext propagation across asynchronous Lambda invocations, fixing trace accuracy for cross-function event flows.
• Built a custom OpenTelemetry collector handling up to 500,000 metrics and trace data points per day.
• Built a CLI tool to automate Lambda instrumentation and AWS infrastructure setup, cutting setup time by up to 40 hours.

stack: Node.js · OpenTelemetry · AWS Lambda · Terraform · Docker · Promscale / TimescaleDB · Grafana