AI · Tooling · Workflow
I push my comfort zone every week — running real product work through agentic editors, model-driven planning, and code I review but don't always type. This page is the meta: what I use, how I use it, and what's shipping next.
Static HTML, CSS, and JS — but the path from idea to deploy runs through an AI-native stack. Each step does one thing well.
Editor where I write, sketch, and prompt. Nimbalyst sessions plan the work; Cursor's edits land in real files.
Reviewer, refactorer, and designer-in-residence. Reads the codebase and writes changes I sign off on.
Source of truth. stage for previews, main for prod. The diff is the contract.
Auto-deploys preview URLs from stage. Quick visual QA before promotion.
Public site at maxwellhandler.com. Static, fast, and boring on purpose.
A working toolkit, not a wishlist. The "exploring" and "on the radar" columns shift quarter to quarter.
Append-only history of what's changed on the site, and what came out of each pass through the loop.
/ai/ describing the AI-assisted build pipeline, daily-driver tools, and what's coming in V2.stage → main flow./assets/ and tightened CSS / JS housekeeping along the way.A small autonomous loop that watches the lending-platform space and proposes site updates when the narrative shifts.
Brave Search pulls fresh signals on competitor moves and category language → Claude distills what changed and why it matters → Claude Code drafts targeted edits to this site (positioning, skill chips, iteration-log entries) → I review and merge. Goal: a portfolio that updates itself when the market does, with a human in the loop on every commit.
Building with AI inside a regulated org, exploring agentic workflows, or just want to swap tooling stacks — happy to talk shop.
San Diego, CA