hey,
I’m AyushiSenior Product Designer, building B2B products from 0→1 through scale — and increasingly, AI-assisted ones.
I like building from 0→1 — and staying long enough to see the system grow.
Senior Product Designer, building B2B products from 0→1 through scale — and increasingly, AI-assisted ones.
I like building from 0→1 — and staying long enough to see the system grow.
Location intelligence, AI-assisted shopping, field operations, and store performance — each asked a different question, but the design challenge was similar: help people see what matters and know what to do next.
Turning location intelligence into a product that helps expansion teams decide where to open next.
Led UX for Lenskart's conversational AI stylist — trust model, voice flows, fit guidance, and AI-assisted checkout.
Unified three roles — brand teams, BD, and independent brokers — into one workflow platform, replacing scattered WhatsApp threads and spreadsheets with structured submission and review.
An alert-first dashboard for area managers overseeing dozens of live stores — surfacing sales, staffing, inventory and operations issues, with compliance cutting across them, each paired with a suggested action.
Across different products, the same hard design problems keep showing up: multi-role workflows, AI that needs to be understandable, and systems that have to surface what matters first.
Different roles rarely need the same interface — but they often need to act on the same underlying record. MarketConnect brought brokers, BD teams, and brand workspaces into one staged property workflow while giving each role the view it needed.
The hard part of AI UX isn’t making the product feel intelligent. It’s helping people understand enough of what it is doing to stay in control. B.ai made recommendations, fit, sizing, voice interaction, and checkout progressively more visible and checkable.
When one person is responsible for many things at once, another dashboard isn’t enough. PerformanceIQ was designed around attention: surface the issue first, show the evidence behind it, then move it into tracked work.
I like understanding why we’re building something before I start designing it. If the problem is still fuzzy, I’ll usually spend more time asking questions before opening Figma.
I’m direct about what I think, but I care more about getting to the right answer than defending my first one. I work best with PMs and engineers who are comfortable challenging each other.
A lot of my work has been in domains I didn’t know before joining. Retail expansion, location intelligence, test automation, AI — I’ve learned to get useful quickly by asking good questions and building a mental model of how the system actually works.
I’ve often owned products as the primary or sole designer, so I’m comfortable sitting with decisions that don’t have a clean design answer. Sometimes the right solution is a new interaction. Sometimes it’s changing the workflow. Sometimes the product should simply do less.
I keep choosing complex B2B products because I like being in rooms where the product, technology and business problem have to be worked through together.
I do my best work with ambitious teammates, difficult problems, and strong trust. The projects below are usually the result of many conversations with product and engineering — not design happening in isolation.