Healthcare · Fintech · AI

I own the whole arc, from research to a live product.

I turn complex, ambiguous problems into business opportunities. The decisions that matter usually get made in the seams between research, design, and code, so I like to stay close to all three.

Builds and shipsI designed and built LendSight's consumer product using AI tools, now live. The bank dashboard I designed is in pilot with a credit union.
AI judgmentI build with AI every day. I'm clear on when to bring it into the product and into my process, and when to keep it out and let my own judgment lead.
Research backedReal user studies sit behind my design work, like a 545-person survey and rounds of user interviews, and some of it is published at CSCW 2025 and in Nature.

Currently working on

Building LendSight, a benchmarking and forecasting tool for community banks, now being piloted with a credit union.
Building mini AI projects to keep pace with the tech and tickle my creative brain. The latest is a CBT companion for the reflection that happens between therapy sessions.
Splitting my weeks between New York and San Francisco, where most of the people I want to build with and learn from are.

Open to new Product Design roles

I started in psychology

I started in counseling psychology, and the hands-on parts shaped me more than the coursework: teaching kids in under-resourced classrooms at the Akanksha Foundation, an internship at Chaitanya Mental Health Center alongside patients living with schizophrenia and bipolar disorder, and a thesis where I ran my own study on resilience in breast cancer patients, through treatment and remission.

What I really learned was how to uncover what a person needs, including the needs they can't name themselves. That turned out to be the same skill design depends on. I was learning to understand user needs long before I ever called anyone a user.

Darshi clustering handwritten notes on a wall

Then I followed the decisions onto a screen

I moved into design because it let me bring that same attention to far more people than I could reach one at a time. My first job was at a design studio where I learned what it takes to build for real users on a real deadline.

At Georgia Tech I focused on healthcare and AI, and the work I'm proudest of was the Hinge Health project, where I designed a women's-health feature end to end, and it landed really well. Alongside it, I did accessibility work at the American Foundation for the Blind and led the research on a WorldCat redesign.

Recognition · Hinge Health

Outstanding Capstone1 of 2 from 74 teams
UX Design AwardsNominated
iF Design AwardNominated

Darshi in graduation regalia in front of the Georgia Tech tower

Now I design and build

At the Financial Services Innovation Lab at Georgia Tech, I designed and built LendSight: a consumer product for comparing banking options, now live, and a bank-pricing dashboard now in pilot with a credit union.

I ran the research, designed the screens and the design system behind them from scratch, and built the front end in v0 and Claude Code, staying close enough to the code that design and build moved together. More on that in how I use AI.

Reading and listening

Currently reading Theo of Golden (Allen Levi).
Just finished Yesteryear (Caro Claire Burke).
Up next is East of Eden (John Steinbeck).

Working through Lenny's Podcast on all things product.
My go-to on runs is Stuff You Should Know.

Off the clock

Six years of Hindustani classical vocal training, five certificate exams.
Chasing the perfect espresso shot and getting better at latte art.
Crafting cold-process soap, a COVID hobby that grew into a small business and an ongoing obsession with gradients.
Dabbling in art, mostly watercolour and digital illustration, with a run of line-drawn flowers lately.

A hand holding a bar of blue-and-white cold-process soap against a leaf

What I believe

There's no substitute for knowing the people you design for.

I spent years working directly with the people I was trying to help, long before I ran my first usability test. It left me convinced that the better you understand a person and what they're really trying to do, the better your odds of building something that fits them.

Most of the work is deciding what to leave out.

The products I work on are dense and high-stakes, and the hard part there is almost never what to add. The most useful thing I can do is decide what someone shouldn't have to deal with right now, and my best work is usually where I took the most away.

Conviction, with the freedom to change course.

In my work I've found I rarely have enough information to feel fully certain, and I've come to believe speed beats perfection. So I move ahead on my best judgment, and when new evidence pushes back on the original idea, I change direction without too much ego about it. Moving quickly on good judgment matters more now than ever, with AI moving as fast as it does.