
Restructured 30+ pages in 12 days
Sole designer and PM. IA, rollout, stakeholder alignment. CEO from day one.
Force-migrated every user. One reverted, then came back. Launch month set 13-month highs in actives and power users.
Read Case StudyI lead design and product for Platform for VC and growth-equity firms to find and evaluate private companies.Go to marketing page
I joined as an intern five years ago and lead its design today. The team has run without a PM for two years, so the product is mine too.
The surface I design and own is a product used by top VC and PE firms
Platform for VC and growth-equity firms to find and evaluate private companies.Go to marketing page is the data platform firms use to invest in private companies. It pulls signals like hiring, web traffic, app downloads and funding into one place, so investors can assess companies that don’t publish their own numbers. Customers include General Catalyst, GIC, Felicis, Greenoaks, Ribbit, Avenir and around 70 more.
All the case studies below share one operating model. I take a platform-scale problem from framing to production, carry the product decisions myself, and ship in weeks. Each is a system: a metric layer, a navigation model, a signal taxonomy, a design system.
Currently building with AI
Synaptic’s new search product. Describe the company you want in plain English instead of stacking filters. I’m the only designer and the product owner: the flow, the marketing website, and the React + Next.js web-app build are mine, built with Codex and Claude Code. Working directly with the CEO from day one.
Go to marketing pageSole designer and PM. Hierarchy, personalisation model, phased rollout.
Recurring requests down 75%. Share of activity 36% → 64% over eight months.
Read Case Study

Sole designer and PM. IA, rollout, stakeholder alignment. CEO from day one.
Force-migrated every user. One reverted, then came back. Launch month set 13-month highs in actives and power users.
Read Case Study
Lead designer and product owner, no PM. Taxonomy, signal logic, five data-ops cycles
Turned an unfiltered people search into an evaluation tool. Firms asked to extend it with their own signals.
Read Case Study
Led design in a 15-person team. Mobile IA, interaction, system architecture.
Mobile actives 9 → 53 in ten months. 51% adoption, the highest the product has seen.
Read Case StudyGo to AppStore page
Sole designer. Tokens, components, dark mode, governance
The platform's first dark mode. Scaled from mobile to web and the Chrome extension.
Read Case StudyI turn noisy, complex domains into tools experts actually use. I do the research, make the product calls, and increasingly build it in code.
I do the research myself and condense it. Two days in the raw data before designing the employee layer, then five cycles with data ops to get 14 signals honest across 700M profiles. The research is where the work happens.
Complex problems rarely need complex answers. A three-tier metric hierarchy, a plain chart where a clever one would slow the read, a company-age line that makes a tenure number mean something. I pick the simpler option and defend it.
For two years the team has run without a PM, so scope, sequencing and the calls sit with me. The cuts are mine to defend: a forced migration with a rollback path, binary signals that hold across firms, the call on what ships now and what waits.
Canonical's product and marketing site are mine, built in Next.js with Codex and Claude Code. I ship in production and close the gap between the mockup and what goes live.
The work runs with a small team: engineers, data, sometimes fifteen people, often the CEO. The product calls are mine, the building is shared.
The same engine works wherever experts drown in data: investing, AI products, scientific tools, music discovery.
Fine arts, then architecture, then product. I taught myself the software each step needed, from AutoCAD 📐 and Revit to Figma, and now I build in code. 👩🏻💻
Good design, to me, is complex work done without friction: fast, smooth, no thinking about the tool. The part I can’t leave alone is the small details, the ones software skips that decide how the whole thing feels. I notice it everywhere: a wardrobe arranged by use, honest brick 🧱 in a brutalist wall, the screen that gets out of your way.
Soft spot for open-source and people building something good. Also, I take my Elaichi chai and music playlist a bit too seriously. ☕️
