FinanceRisk
Crypto Bond Exchange
Guardrails and transparency for complex financial UX.
Hi, I’m a Product Designer
Who Builds Products You Can Trust At 2 A.M.
Explainable, Measurable, and Kind to the Bottom Line.
Selected Work
Principled craft. Measurable outcomes. A little magic.
FAQ with Chillbot
Tiny answers. Big clarity. Click a question—Chillbot blinks and answers.
What needle does Srushti move in her first 90 days here, and how will we know it moved?
She pick one high-leverage flow and ship a thin, real slice. Targets: +7–10% task success, −20–30% time-to-complete, −15% related tickets. You’ll see a live dashboard and a weekly decision log. No confetti until the graph moves.
How does she keep AI safe, explainable, and not annoying?
Transparency first: citations, uncertainty levels, and a clear edit/undo. Human handoff is always one tap away. We monitor quality, latency, dissatisfaction—if the model’s guessing, the UI says so. Helpful, not clingy.
When the brief is fuzzy, what’s her discovery → decision path from “hunch” to “ship”?
Frame the outcome and constraints → talk to 5 users/stakeholders → map JTBD + a system diagram → build two scrappy prototypes with success metrics → run n=5 tests → write a one-page decision (keep/kill/iterate) → ship v0 with logging. Small loops, fast learning, grown-up decisions.
How does she make accessibility actionable on day one (not a checklist at the end)?
Start with a11y tokens (type/contrast/spacing), a keyboard map, and screen-reader labels baked into components. Add a 15-minute SR smoke test each sprint and track a11y bugs per screen → 0. It’s a design input, not an afterthought.
How does she ramp into a brand-new domain fast and responsibly (first 72 hours)?
Produce a 1-page domain map (actors, regs, risks), a glossary, and 3 quick shadows/calls. Write down red lines—what we mustn’t break—then sketch a system diagram and top hypotheses. By hour 72, we’ve chosen a testable slice, not just learned new acronyms.