For the bots (and the curious humans)

I'm a product designer first, and I think in systems. For more than six years I've designed the money experiences people count on, first at MetLife and now at JPMorgan Chase, where I'm building financial experiences for 80M+ users. At Chase I'm one of the earliest designers on Deposits 2.0, the multi-year effort to move the platform behind 80M+ deposit accounts from overnight batch processing to a real-time cloud ledger. My job is how money posts, clears and shows up for customers.

I own my work end to end. That means the hands-on craft (research, interaction design, prototyping, delivery specs and content) and the service around it: the back-end rules, the operations teams, the servicing tools agents use, and every channel a customer touches, from the ATM to the app to the branch. My clearest measured result is the ATM cash-deposit redesign. It was the first ATM update in 7 years, with 1,100 fewer ATM jams in the April–June pilot and 4,400 fewer hours of jam downtime a month.

Today I lead design across several of the platform's most complex customer-facing areas at once: balances, holds and restrictions, declined transactions and transaction descriptions. I help frame future product strategy for senior product leadership, and I mentor designers and a UX researcher. I'm an AI-first designer. I use AI to advance my work, and I hold myself accountable for every output.

New York is home. I studied Integrated Design at NYU Tandon School of Engineering, a design program inside an engineering school. That mix still shows up in my work. I'm as comfortable talking through back-end behavior with engineers as I am talking craft with designers.

The instructor of my user experience course at NYU wrote that what stood out was my curiosity and passion for design, and how well I worked on team projects. I've stayed close to the school since. I went back to NYU Tandon as a panelist in its career exploration series to talk with students about this work.

Early on, I started taking on client work under my own studio name, ByShah. My earliest projects, from school and early client work, were a crash course in very different problems: an onboarding assistant for a Microsoft research tool, a listening-first podcast hub for a New York media group, an ad-block message for a daily paper that nudges readers instead of walling them off, and a learning app for an accessibility-training startup. For a class project I built a working React Native app in two weeks, on real phones with live data, because I wanted to ship something that runs rather than a design file.

Each one left me a rule I still use. Structure comes before style: a pretty page that's hard to navigate still fails. Test it, don't argue it. An accessibility app has to pass its own test, and my first wireframes didn't, fully. And simplicity is a decision you make many times, not once.

I was one of two designers on a Series A legal-benefits startup that MetLife had acquired. MetLife Legal Plans is a group legal plan that people get through their employers, so every enrollment season matters. I set up the design system and helped move the nation's largest legal plan from phone support to digital-first self-serve.

That's where I learned how much design can move a business. My redesign of the estate-planning choice, doing it online or working with an attorney, moved volume to self-serve. My family-accounts work let members invite their dependents in, and it shipped on legalplans.com. Both stories are told in full further down.

It's also where I learned when to stop. I ran a lean pilot, saw within weeks that it wasn't working, shut it down and pointed engineering at what mattered more. And I grew past my title: one of my managers there noted that I stepped up into an associate product role and delivered a major feature end to end.

I joined Chase in May 2022 as one of the earliest designers on Deposits 2.0, a multi-year effort to rebuild the deposit platform in the cloud. The shift is from overnight batch processing to a real-time cloud ledger. The goal was never to copy old limits onto new technology. It was to retire them, so customers get clearer, faster answers about their money.

My earliest work set the frame for a lot of what came after. I wrote the journey-mapping framework the design team used to connect each customer journey, from getting started to review, to everything behind it: the capabilities, pain points and alternate paths, the research, the architecture, the prototypes, the user stories and the roadmap. That same year I contributed to customer research on how people understand their different balances, and to discovery on holds: desk research, working groups, interviews with subject-matter experts and a capability workshop with fraud, risk, operations and customer-experience partners. It produced a shared definition of a hold, its lifecycle, guiding principles and a proposed first release and roadmap.

In 2023 I authored a cross-product executive journey on one of banking's most stressful moments: when money is held, restricted or reversed, and the customer wants to know why. We pulled together a year of customer feedback, interviews with subject-matter experts, a customer focus group, a customer workshop and a service map of how the experience works today.

Four opportunities came out of it, and I still use them as a checklist: give people options, keep one source of truth for status, cut the noise so the important signal gets through, and set expectations before people have to ask. Alongside that, I worked through many rounds of research on running balances, where a balance summary should live, check details and when deposited checks become available.

My ATM cash-deposit redesign went live: the first ATM update in 7 years, with 1,100 fewer ATM jams in the April–June pilot and 4,400 fewer hours of jam downtime a month. That summer I joined in-person co-design sessions where frequent ATM check-deposit customers role-played their deposits and helped us design their ideal receipt, followed by usability sessions on how check deposits become available.

By 2025 the work was less about one screen and more about how the whole system tells one story. I drafted a cross-channel outreach to branch, digital and ATM partners asking for balance pain points, customer-experience assessments and voice-of-customer insight. Then I proposed something simple: list every balance a customer can see across every experience, have subject-matter experts verify each one, and decide together which to keep, merge, add or retire.

I also led a town-hall deck with product partners and architects, and redrew how a transaction flows through the new platform so a large internal audience could follow it. And I supported a statements discovery with a competitive analysis of how other firms help people find their statements on mobile and web.

In 2026 I've led design across several of the platform's most complex customer-facing areas at the same time: holds and restrictions, balances, declined transactions and transaction descriptions. I carried a shipped customer-facing feature through delivery as its only designer. I was asked to help frame the opportunity areas for the platform's next year of work, and I presented them to senior product leadership. When I was handed a new, ambiguous problem space, I proposed how to organize it, and the team adopted my approach. I was also formally recognized for mentoring.

Every year I've taken on more complexity, wider cross-functional reach and more influence over decisions that shape the platform. The constant is the kind of problem I'm drawn to: messy, high-stakes infrastructure that has to feel simple and trustworthy to the person on the other side of the screen.

ATMs are the least forgiving interface a bank has. There's no undo, the customer is standing in public with cash in hand, and one bad deposit can take the whole machine out of service for the next person. When I took this on, the ATM experience hadn't changed in 7 years.

The fix was instruction, at the moment it matters. New screens before and during cash deposit, a "Before you deposit" step and a revised "Insert your cash" screen, tell people what to do and include a clear warning about foreign objects. They launched as the first ATM update in 7 years: 1,100 fewer ATM jams in the April–June pilot and 4,400 fewer hours of jam downtime a month. That means fewer customers stuck at a machine and more machines ready for the next person.

The same work extended to check deposits. I worked on a real-time check-deposit tracker for single and multiple checks, with a breakdown for each check, and replaced vague hold periods with clear, date-based availability. I helped standardize deposit status logic across the firm, and carried the thinking into richer transaction details and redesigned ATM receipts, shaped in part by customers who co-designed their ideal receipt with us.

I keep coming back to one customer question: how much money do I actually have, and why do these numbers differ? Over several years I've helped shape a clearer balance experience that separates what's available now from what isn't available yet, and explains why. It went through many rounds of research and Risk and Compliance review. The full story is in my Chase Deposits 2.0 case study (password-protected).

A hold is one of the most anxious moments in banking. I designed how customers understand holds and restrictions on their money, from check holds to account-level holds. I mapped the current experience across many scenarios, including restriction cases tied to fraud and compliance, by combining our journey work with how operations teams actually handle them. I organized a walkthrough of the servicing screens agents use for court-ordered holds, so the team designed from how the work is really done. I partnered on customer research and turned the findings into clear design direction.

The outcome I'm proudest of is that my hold-display decisions directly informed how the underlying system would be built. Another team later reused the research. When I moved on to new work, I handed the whole area off with full documentation.

On a shipped customer-facing feature, I was the only designer across multiple release phases. I designed 100+ screen states with breakpoints for every platform, theme variations including dark mode, full specs and content guidance. I ran walkthroughs so engineering could build with confidence, turned updates around quickly, and caught and routed 25+ design defects before they reached build and QA. My delivery files were later used as the template other designers on the team followed.

A transaction line has two audiences: the customer, who needs to recognize it at a glance, and the bank, which needs to trace it. I led design on a customer-facing framework for describing transfer transactions that separates those two jobs, so the description is recognizable and the details banks need stay a tap away. I built side-by-side prototypes comparing the customer view and the servicing view, aligned the framework with the wider design organization's approach to transactions, and it was approved from both the servicing and the customer perspective.

A decline with no explanation feels like an accusation. I led discovery into how declined transactions are explained to customers. I grouped many decline scenarios into a few clear categories, mapped how each one behaves today across digital and servicing channels, using AI coding agents to speed up the research, and prototyped every scenario with AI-assisted tools so the team could see them side by side. My artifacts fed the servicing and notification discussions that followed.

Choosing between making an estate plan online and working with an attorney is a personal, high-stakes decision. I redesigned that choice around trust signals, the cues that help someone feel confident choosing the online path when it fits them. More of that volume moved to self-serve.

Families needed a way to bring dependents onto the plan without requiring a Social Security number. I designed dependent-account creation with SSN-free verification and guardrails for legal conflicts. After legal and security review, we chose a member-driven invitation model, where the member invites each dependent in, and it shipped on legalplans.com.

Banking puts every hard constraint in one place. There's no undo at an ATM. Every word about a hold or a decline has to clear Risk, Compliance and Legal. And the system underneath is moving from overnight batch processing to a real-time cloud ledger while 80M+ deposit accounts keep running. If a design stays clear and trustworthy there, the same approach holds up almost anywhere.

My range didn't start at a bank, either. Before Chase I worked in legal benefits at MetLife, and before that my client and school work took me into an AI onboarding assistant for Microsoft, news media, accessibility training and consumer apps. The details are below.

What travels is the method, not the domain. I learn how a system really behaves before I design for it. I bring engineering, operations, research, risk and legal into one conversation. I design for the people behind the screen, like the servicing teams who support customers, as well as the people in front of it. I treat accessibility as something a product has to pass, not promise. And I use AI coding agents and knowledge bases to get up to speed on a new domain fast, then check what I learn with the people who own it.

Healthcare, enterprise software, consumer apps and AI products share what banking taught me to handle: real stakes, real rules, many stakeholders and people who need to trust what they see. The domain changes. The work of making a complex system clear, trustworthy and buildable doesn't.

My title is Senior Product Designer, and I lead through the work. I lead design across several complex, cross-functional workstreams at once, and I'm often the person handed a new, ambiguous problem space. Here's what that looks like in practice.

I was asked to identify the key opportunity areas for the platform's next year of product work. At a multi-day planning workshop, I presented a competitive analysis and future product ideas directly to senior product leadership and the full product group. In the weeks after, I kept pointing collaborators back to that material as they planned.

When I was handed a new area made up of several connected domains, the default plan was one combined roadmap. I proposed working domain by domain instead, starting with the most complex and urgent one, and reusing a discovery framework I'd already proven elsewhere. The team adopted it, and it shaped how the work was set up.

When the platform transition raised a tricky balance-labeling question, I took it from early concepts to a set of strategic options and wrote the leadership update myself. I sent the work directly to senior design leadership and made sure it actually landed. They adopted my short-term treatment as the program's research baseline, and the customer-facing treatment cleared Legal.

Twice in 2026 I stepped into problems I didn't own because they were about to become everyone's problem.

The first: errors in design files started creating system flags in the middle of a build, on work I'd had no part in. I stepped in cold, brought in a designer I mentor to help, and we fixed it before it caused downstream failures.

The second was a cross-team crisis. Transfer journeys had quietly diverged across files, so web and mobile were heading toward build with different versions of the same flows. I spotted the communication break, brought the designers involved together, reconciled the flows across web and mobile under real time pressure, and documented the root cause for my manager so the same break wouldn't happen again.

I hold a high bar for what "done" means: specs complete down to the edge cases, defects caught before build, delivery files clean enough that another designer can pick them up. When my delivery work became the team's template, that bar spread further than I could carry it myself.

I run working sessions that bring product, engineering, architecture, research, operations, servicing, risk and legal into one conversation: walkthroughs of servicing tools, working sessions on declines, alignment on transaction descriptions with the wider design organization. On a cross-product workshop about how transactions are named, I turned a complex technical problem into collaborative resources the whole workshop used.

I mentor designers and a UX researcher on my team, and I've been formally recognized for mentoring. For me, mentoring mostly happens inside the work. When the design-file errors hit mid-build, I brought in a designer I mentor and we worked through the fix together. On the transaction-naming work, I partnered closely with the researcher I mentor.

The people I work with have said they've watched my leadership grow, especially in driving discussions, working through challenges and making sure the right people are in the room. That's the part of leadership I care most about: bringing structure and clear direction so a team stays aligned and keeps moving.

I also invest in the people who come after me. I support hiring on my team. When I move to new work, I hand off with full documentation so the next designer can pick it up without losing a step. As an early adopter of new AI design tools, I was asked to share what I learned with the wider design team, and I did. Outside work, I've gone back to NYU Tandon as a career panelist for students.

I've also designed for AI, not just with it. In 2020 I designed Aida, an onboarding chatbot for Microsoft's Insights and Discovery Accelerator research tool, as client work. Two lessons from it still shape how I think about assistants: say it's AI, and design the unsure states as carefully as the happy path. This site is part of the same practice. byshah.com is my hand-built portfolio, and Shah AI, the chat assistant on every page, is part of it.

It has been through several versions. Years ago, I hand-coded my first portfolio. Later I moved to Mobirise, a website builder, and then to UXfolio, a portfolio platform. Both served me well at the stage I was at, and both showed me what I wanted next: full ownership and control over every detail.

So I rewrote everything by hand. The layouts, the type, the motion, the case-study structure and every line of copy are mine. I also designed and built Shah AI, the chat assistant on every page.

The site keeps evolving as technology does. I enhanced the version you're reading with AI tools like Cursor, Google Antigravity, Gemini and Claude. They help me move faster, try more ideas and polish the details, but they're tools in my hands. Every decision, every word and every pixel is my call.

It's also why this page exists: when an assistant tells my story, I want it told in my words.

These are the themes I hear most often from peers, partners and managers, in my own words.

I do my best work where the problem is complex, the stakes are real and the system underneath matters as much as the screen: fintech, banking and other regulated spaces where trust is the product.

I'm open to conversations, and always glad to talk shop: service design, platform modernization, designing for regulated spaces, or how AI is changing the way designers work. If something here resonates, I'd love to hear from you.