Work & Products

Protected. Enter the password to view this case study.

Baavy preview

Problem

Researching viral content eats hours: hopping between tabs, and IG/TikTok give you no way to sort a feed by what actually performed.

What I built

Baavy, a research tool that ranks any Instagram or TikTok feed by engagement, so the top-performing posts surface in seconds. It’s a web app plus a Chrome extension: the extension reads public data through your own browser session (no third-party login, no platform rate limits), and the app turns it into a sortable, exportable board with ~50 ready-made niches.

Approach

  • Designed and built it end-to-end: product UX, the extension, and the marketing site.
  • Bahasa-first and IDR-priced for the Indonesian creator market.
  • A privacy-conscious scraping model: the work runs in the user’s own session, so there’s no shared account to get blocked.

Status

In closed beta at baavy.id.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.

Hikery preview

Context

Hikery.co is an outdoor travel startup (equipment rental and open/private trips). In 2022 they were validating product–market fit with college hikers around Greater Bandung.

Goals

  • Business: grow brand awareness and repeat visitors
  • Conversion: more bookings through the site

Challenge

Early-stage startup with little quantitative social proof. The page had to explain two services clearly (trips vs rentals) while still feeling credible.

Process

  1. Audited the existing desktop landing page
  2. Wireframed a clearer narrative arc
  3. Designed hero, social proof, USP, services, feedback form, and footer
  4. Applied an F-pattern layout for scannable copy on hero headlines

Design decisions

  • Hero: benefit-led headline and supporting visual on the left for scanning
  • Social proof: real testimonials instead of inflated stats while PMF was still open
  • Services: split trip booking and gear rental with live inventory cues where possible
  • Feedback form: client-requested loop for qualitative learning pre-scale

Outcome

Shipped as my first public UI/UX case study. It set the tone for later founder-facing landing work: Sharehub, MilikiRumah, and beyond. Read the long version on Medium.

Protected. Enter the password to view this case study.

Masjid Review Indonesia preview

Problem

Finding a good mosque in an unfamiliar area (its facilities, atmosphere, and real condition) usually comes down to guesswork or word of mouth.

What I built

Masjid Review Indonesia, a community platform for discovering the best mosques in your city, ranked by real experiences from the people who pray there. Worshippers share what a mosque is actually like, so newcomers and travellers can choose with confidence.

Approach

  • Designed and built it end-to-end, from concept to shipped product.
  • Put community experience at the center: reviews from real worshippers rather than top-down listings.
  • Kept it simple and fast, so finding a nearby mosque takes seconds.

Status

Live at masjidreview.id.

Protected. Enter the password to view this case study.

Miliki Rumah preview

Problem

Freelancers and small-business owners in Indonesia struggled to access credible rent-to-own paths into home ownership.

What I delivered

  • Web application UX for the rent-to-own journey
  • Landing page narrative for trust and eligibility
  • Brand and UI direction aligned to a regulated, high-trust category

Result

The product launched publicly. Shortly after launch the business announced a $50m fund, a strong signal that the proposition and execution resonated in market.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.

Context

Realcreator is a whitelabel booking product: media companies run it under their own brand so their own customers can book their services through it. So this flow has two audiences to keep happy: the customer moving through it, and the brand shipping it as if they’d built it themselves.

The first time I clicked through the existing flow, though, it felt less like booking a service and more like a job interview I hadn’t prepared for. Lots of questions up front, very little given back. The bones were solid. It just kept asking before it had earned the right to.

The problem

The flow lost people at almost every step: a yes/no question and a sidebar stepper eating a third of the screen at entry, a calendar that showed date, time, and operator at once with no sense of what was free, and a dead-end confirmation with copy bugs and mismatched totals. Scheduling was where most people quietly gave up.

What I did

I gave the whole thing one rule to live by: one decision per screen. Then I re-ordered the questions so each is easy by the time you reach it. Three of them (suggested time, operator, itemized bill) moved off the main screens into modals that only appear once the flow has enough context to make them trivial.

Highlights:

  • Entry asks for one thing, your email, and greets returning customers with “Welcome back.”
  • Schedule shows availability first: full days struck out, capacity dots, slots grouped by morning/afternoon/evening, timezone stated out loud.
  • Services lead with a “For You” default; the running total tucks into a bar you open on demand instead of hovering the whole time.
  • Confirmation ends with somewhere to go: self-serve cancel, reschedule, and “Make another booking” for the regulars.

Outcome

A calm, whitelabel booking wizard any media company on Realcreator can ship under its own name, with the step that used to lose the most people rebuilt around availability you can finally see.

Protected. Enter the password to view this case study.

Roots preview

Problem

Getting Gen Z to stick with real-world sustainable habits is hard: one-off pledges don’t build a routine.

What I built

Roots, a gamified eco-habit app that runs where its users already are: WhatsApp. Each morning it sends a daily mission; you reply with a photo, an AI vision model verifies it, and you earn points redeemable for local-brand vouchers. The web dashboard shows streaks, impact (CO₂), and a leaderboard.

Approach

  • Designed and built it end-to-end: the WhatsApp bot flow, the AI photo-verification pipeline, and the progress dashboard.
  • Chose WhatsApp over a native app to remove the download barrier for a habit-forming audience.
  • AI verification with a confidence threshold, falling back to human review, so the points stay trustworthy.

Status

In pilot at rootsapp.id.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.

Service Subscribers preview

Problem

Australian SMBs sold services ad hoc but lacked a lightweight CRM to convert one-time buyers into recurring subscribers.

What I delivered

  • CRM-oriented product UX
  • Marketing site explaining the subscription value prop
  • Brand and UI cohesion across app and landing

Result

Shipped and live at servicesubscriber.com.

Sharehub preview

Problem

New Zealanders needed a trustworthy way to rent everyday gear from people nearby, and owners needed simple flows to list idle items.

What I delivered

  • Product UX for the rental marketplace
  • Landing page and brand-aligned marketing surfaces
  • Visual system carried through the web app’s key flows

Result

The platform went live. Founder Sepideh Zoey has publicly endorsed the collaboration.

Skill Assessment App preview

Problem

Bootcamp participants filled self-reflection checklists after class, but the process was subjective. People could mark skills as “mastered” without matching learning goals, which hurt both confidence and mentor oversight.

What I did

  • Kickoff with the client on expectations and how classes ran today
  • Classroom observation plus interviews with participants
  • Synthesized pain points into needs for students and mentors
  • User flows, sitemap, lo-fi wireframes, hi-fi UI, and an interactive prototype

The client chose mobile as the primary surface after we aligned on the problem.

Solution

A skill-assessment app where quizzes map to published learning standards, so progress is measured against objective criteria instead of gut feel. Core journey: sign in → pick assessment → complete quiz → see pass/fail with clear next steps.

Outcome

MVP design delivered in one week (20–27 Jul 2022). Usability testing was run by the client under NDA, so public write-ups stop at the prototype stage. Read the case study on Medium.

Protected. Enter the password to view this case study.

Teman Tanam preview

Problem

Keeping plants alive means remembering watering and care routines that are easy to forget, and most tools ask you to download yet another app first.

What I built

Teman Tanam, a plant-care companion that runs right in the browser, no download required. Track each of your plants and get care reminders delivered on WhatsApp, so nothing slips.

Approach

  • Designed and built it end-to-end as a lightweight, install-free utility.
  • Met people where they already are: reminders over WhatsApp instead of push notifications from a new app.
  • Kept the experience friendly and low-effort, so caring for plants stays a small daily habit.

Status

Live at temantanam.app.

Protected. Enter the password to view this case study.

Protected. Enter the password to view this case study.