Mobitino
Payments, people and places — in one confident app.
Mobitino is a mobile app that fuses UPI payments with hyperlocal discovery. In one place, people pay and request money, find nearby stores and people, follow a community buzz feed, and chat — with payment woven right into the conversation. I designed the end-to-end mobile experience, from first-run onboarding to the pay flow, discovery, and profile.
The hard part
Bundling payments, discovery, chat and a community feed into one app risks a bloated, confusing product. The challenge was to keep payments fast, secure and trustworthy — the reason people open the app — while making local discovery and social feel effortless rather than distracting. Onboarding also had to earn trust quickly before asking for money and location access.
What the research told me
How I worked
Map the jobs
I separated the core jobs — pay, discover, connect — so the app could hold all three without confusion.
Anchor on payments
I made the Ask / Pay action the heart of the home screen, reachable in a single tap.
Layer discovery & social
I wove nearby stores, people and a community buzz feed around payments without crowding them.
Prototype & test
I prototyped the pay and onboarding flows and tested until each felt fast and reassuring.
The shipped design
Mobitino ships as one app for money, people and places. Payments sit one tap away, a hyperlocal feed surfaces nearby stores and people, and chat carries payment inside the conversation — so sending money feels as natural as sending a message.
- A one-tap Ask / Pay action at the heart of the app, with UPI and QR
- A hyperlocal explore feed for nearby stores, people and community buzz
- Chat with payment built directly into the conversation
- A clean profile with QR, linked accounts and transaction history
The impact
Pickl.AI
Bridging the gap between learning data science and doing it.
Pickl.AI is the educational vertical of TransOrg Analytics, built to close the distance between theory and applied data science. The platform spans a large course catalogue, a flagship Job Guarantee Program, live events, city-specific landing pages, and verifiable certificates — all of which had to feel like one coherent, trustworthy product on both web and mobile.
The hard part
A learner’s decision to invest in a course is high-stakes and research-heavy. The platform had to earn trust fast — clear curricula, transparent pricing, outcomes and alumni proof — while covering a sprawling scope of pages (courses, program, events, careers, certificates) without ever feeling generic. Every template also had to work as hard on a phone as on desktop, since most learners browse on mobile first.
What the research told me
How I worked
Map the scope
I built the information architecture across courses, program, events, careers and certificates so nothing felt bolted on.
Design mobile-first
I designed each template on mobile first, then scaled it up to desktop, since most learners start on a phone.
Build for trust
I foregrounded curricula, pricing, outcomes and alumni proof so learners could decide with confidence.
Systemise it
I created a design system so a large, growing catalogue stays consistent across every page and platform.
The system behind it
Before design, I built a shared visual language — the Pickl.AI colour palette, Mulish type scale, and a component library — so a large, fast-growing catalogue would stay consistent across every page and both platforms.
The shipped design
The shipped platform makes learning data science feel credible and effortless — clear course pages, a persuasive Job Guarantee Program, verifiable certificates, and a mobile experience that carries the full journey in one hand.
- Course listing and detail pages built around curriculum, outcomes and transparent pricing
- A flagship Job Guarantee Program page designed to convert serious learners
- Verifiable certificates and a public verification page for employers
- One responsive system spanning web and a mobile-first experience
And the same experience, mobile-first — every key journey rebuilt for the phone.
The impact
Artcab
Great films start with the right people, not a job board.
Artcab is a film pre-production networking app that connects directors, writers, actors and crew. Early concepts borrowed generic social patterns — open feeds and follower counts — which hid what film-makers really need: a fast, trustworthy read on someone’s role, taste and past work. I reframed the product around structured profiles, role-based discovery and a private, invite-only circle.
The hard part
Creative networking is a trust problem, not a volume problem. Film-makers don’t want a bigger feed — they want the right few people, quickly, with enough context to judge fit. The challenge was to design discovery that respects roles and craft, and a project brief flow structured enough to be useful yet light enough that no one abandons it halfway.
What the research told me
How I worked
Learn the workflow
I interviewed film-makers to understand how crews really form — who they trust, and what they need to know before saying yes.
Model roles & profiles
I designed a profile that carries up to three roles, taste and past work, so fit reads at a glance.
Build the private circle
I structured discovery by role and a private circle, so networking feels curated and trusted, not public and noisy.
Brief a project
I shaped a short, guided brief — type, stage, budget — that captures the essentials without becoming a chore.
The shipped design
The final app treats networking as a considered act. Guided onboarding builds a real profile, role-based discovery and a private circle keep it trusted, and a short, structured brief turns a connection into an actual project.
- Guided onboarding — photo, up to three roles, cinema taste and genres
- Role-based discovery with rich, scannable profiles and filters
- A private circle that keeps networking curated and trusted
- A short project brief — type, stage and budget — plus built-in chat
The impact
Uncarbon
Clear enough to sell a buyer in ninety seconds, deep enough to hold an analyst for ninety minutes.
Uncarbon is an AI-powered ESG and carbon accounting platform that helps enterprises measure, reduce, and report emissions. The raw material is messy — inconsistent data scattered across a dozen systems — and the earlier tooling simply dumped it on screen as tables. I designed the brand and the product as one, so the story stays intact from the first pitch to the daily dashboard.
What I owned
I designed Uncarbon end to end, from a blank canvas. That meant creating the logo and brand identity, running stakeholder collaboration and requirements gathering to define what the platform had to do, and then shaping every core flow — data collection, analysis, reporting, and admin. Throughout, I worked in close coordination with developers to keep the build faithful to the design and shippable at scale.
The hard part
ESG data is huge, contested, and rarely tidy. I had to make it feel trustworthy and act-on-able without flattening it — and serve two very different people, a curious executive and a rigorous analyst, inside one interface instead of building two products that drift apart.
What the research told me
How I worked
Separate the jobs
I interviewed executives, analysts, and auditors to untangle what each actually needed from the same numbers.
Model the mess
I mapped how a dozen fragmented sources could reconcile into one trustworthy view before designing anything.
Build brand + product
I designed them together — a calm palette, a disciplined chart set, and dashboards that layer detail.
Cut what didn’t earn it
I tested with real analysts and removed every chart that didn’t lead to a decision.
From rough to right
Before any pixels, I mapped the platform with the team — flows, screens, and data structures for onboarding clients, adding data, and managing facilities and users. These early wireframes aligned stakeholders and developers on scope before design began.
The shipped design
The shipped product organises all that data around decisions, not tables. A calm dashboard leads with the headline, progressive disclosure holds the analyst-grade depth, and one visual language — built on a documented design system — runs unbroken from the marketing site into the app.
- A disciplined chart set that reads at a glance and survives close inspection
- Progressive disclosure — headline first, the full working on demand
- Reconciled views that turn a dozen sources into one narrative
- A design system and brand that never visibly seam apart
The impact
Healthcare knowledge base
The answer already exists — in a closed ticket nobody can find.
This is an AI-powered knowledge base for a healthcare software support team handling thousands of tickets across a dozen medical products. Every resolved ticket held a real answer, but it stayed locked in a closed case. I designed a platform that turns those tickets into structured, searchable articles — with AI drafting the first version and support experts reviewing before anything publishes.
The hard part
A knowledge base lives or dies on trust. In healthcare support, a stale or wrong answer isn’t just annoying — it can misguide a clinic. AI could draft articles fast, but nothing could publish unchecked. The challenge was a system fast enough to keep up with ticket volume, yet safe enough that a human expert signs off on every article, with duplicates caught before they multiply.
What the research told me
How I worked
Learn the support workflow
I interviewed agents and mapped how tickets get resolved to see where knowledge was being lost.
Design the AI-to-human flow
I shaped how AI drafts an article from a ticket and hands it to an expert for review.
Build in trust signals
Confidence scores, similarity checks, and validation states let reviewers approve with confidence.
Make it findable
A clean knowledge base with search, categories, and an AI assistant so answers surface fast.
From rough to right
I iterated on the review and editor layouts — how to show AI confidence, similar-article matches, and validation checks without burying the article itself. Small, well-placed trust signals let a reviewer approve or reject in seconds. The design system kept every screen consistent across the platform.
The shipped design
The platform ships as an AI-powered knowledge base with a human in the loop. AI drafts articles from resolved tickets, experts review with full context, and a clean, searchable library — plus an AI assistant — makes answers easy to find across every medical product.
- A searchable, filterable knowledge base across all products
- An AI assistant for natural-language answers
- Analytics on usage, search behaviour and reuse impact
- Admin controls for AI thresholds, categories, users and access
The impact
AI Voice Agent
Never miss a patient call — even when the front desk is full.
PA is a healthcare AI voice-agent platform. It answers inbound patient calls, handles booking, rescheduling and cancellations, and writes the outcome directly into the practice’s EHR — so front-desk staff can focus on in-person care. I designed the operations product: the dashboards, monitoring, and configuration that let a CGM operations team run the agent across dozens of clinics with confidence.
The hard part
Handing patient calls to an AI is a trust problem, especially in healthcare. Operations teams need confidence that the agent captures intent correctly, books the right slot, and escalates when it should. The design challenge was making an invisible, autonomous AI conversation fully observable — every intent, booking, handoff and EHR write legible, auditable, and controllable from one dashboard.
What the research told me
How I worked
Learn the operations job
I interviewed CGM operations teams to understand how they’d supervise an AI taking real patient calls.
Map the AI call lifecycle
I blueprinted every automated step — answer, intent detection, booking, escalation, EHR write — to define what teams must see.
Design for observability
I surfaced real-time monitoring, per-call AI transcripts, and an audit trail so no agent action is ever a black box.
Make onboarding safe
I designed a guided practice registration and configuration flow so each clinic goes live without misfires.
From rough to right
I iterated on the monitoring and configuration layouts — how to surface live call outcomes, handoffs, and usage without overwhelming an operations lead, and how to make practice onboarding safe on the first try. A dedicated design system kept every screen consistent across a large, data-heavy platform.
The shipped design
The platform pairs an autonomous AI voice agent with the operational controls to trust it. The agent answers calls, understands intent, books and syncs to the EHR on its own; the workspace makes all of that automation visible — guided onboarding, real-time monitoring, usage analytics, and a full audit trail in one consistent, data-dense product.
- An AI voice agent that answers, understands intent, and books 24/7
- Automatic EHR sync so every booking lands in the practice system
- Real-time monitoring with AI handoff and escalation alerts
- Guided practice onboarding, role-based access, and a complete audit trail
Built to launch
Currently in final pre-launch, going live in the coming weeks. The platform is designed around AI-driven automation that removes the front-desk phone bottleneck — here’s what it does and the value it’s built to deliver.
Free the front desk
Automating routine appointment calls lets staff focus on in-person patient care instead of the phone.
Capture every opportunity
Answering after hours and at peak times is built to reduce missed calls and lost bookings.
Scale with confidence
Guided onboarding, observability and audit trails let one team run the agent across many clinics safely.