I turn emerging AI capabilities into software people can actually use — fast enough to learn, solid enough to keep.
I am a Toronto-based software engineer working where AI agents, developer experience, and small, sharp products overlap. My favorite projects remove a tedious loop, expose a hidden cost, or make a complex system feel obvious.
|
Speedtest.net for language models. Benchmark inference latency from any MCP client—with zero telemetry.
|
One cost ledger for every coding agent. See what autonomous development actually costs across CLIs.
|
|
A discovery layer for AI agents. Find and share useful tools, APIs, models, and workflows.
|
A reality check for AI-generated code. Catch security issues, anti-patterns, and maintenance debt before they ship.
|
|
API docs, compressed into action. Paste an endpoint and get the curl commands you need.
|
A living view of model performance. Follow the fast-moving AI model landscape without drowning in release noise.
|
find friction → build the smallest useful thing → put it in real hands
↑ ↓
└──────────── measure, learn, simplify ─────────────┘
I optimize for:
- Useful over theatrical — the best demo becomes a dependable daily tool.
- Observable over mysterious — costs, failures, and tradeoffs should be visible.
- Small surface area — fewer moving parts, clearer ownership, faster iteration.
- Shipping as research — working software teaches more than another planning document.
building:
- agents that can use tools safely across files, browsers, terminals, and chat
- infrastructure for model evaluation, cost visibility, and reliable automation
- focused products that turn one painful workflow into one satisfying action
exploring:
- memory and deterministic conflict resolution for local-first agents
- evals that measure whether an AI workflow is useful, not merely impressive
- security boundaries for software that can act on a user's behalfCore: Python · TypeScript · JavaScript · Go · SQL
Product: React · Next.js · Astro · Tailwind · FastAPI · Node.js
Systems: PostgreSQL · Redis · Docker · GitHub Actions · AWS
AI: MCP · tool use · evals · retrieval · browser automation · LLM APIs