01The problem
Technical-analysis tools are either toy-simple or a wall of indicators only a pro can read. There was room for something in between: a clean app that surfaces one clear momentum signal and lets a regular investor compare stocks side by side.
02What I built
The product and front end for a web app built around a proprietary momentum indicator and a diverging target-range gauge. Users keep watchlists, open deep per-stock views, and line several stocks up on a synced comparison canvas.
- A momentum indicator and target-range gauge as the core read.
- A comparison canvas with a synced crosshair across multiple charts.
- A planned AI concierge that explains the method and helps with the app, under a hard non-advisory ruleset.
03The interesting constraint
Financial software lives inside real legal lines. A concierge that gives personalized buy or sell advice would cross into regulated territory. So the ruleset came first: the assistant is scoped to help you use the tool and understand the method, and it refuses to recommend trades. Defining those guardrails precisely was as much of the product work as the charts.
Why this matters for an AI PM role
This is product requirements for a technically and legally complex domain: defining a ruleset, edge cases, and a hard non-advisory boundary, then directing the front-end build against it. Defining rulesets, acceptance criteria, and failure modes is the core of the work.
04My role
Product, UX, design, and front end (React + Vite), built with Claude Code alongside a partner who owns the backend and deploy. I own how the product looks, reads, and behaves.