PORTFOLIO / AI/ML PRODUCT / BEAUTY-TECH
AI/ML Product
B2B + B2C
Design System
ROLE
Product + Brand + Graphic Designer & Video Editor
TARGET AUDIENCE
B2B partners & companies, B2C
users (aged 15–42+)
SCOPE
B2B Dashboard + Ambassador Program + Website/Web App + Mobile App + Plugin + Brand & Marketing Design
200,000+
Customer & Expert Profiles
Proprietary dataset powering the AI engine
2+
Paying B2B Clients Onboarded
Plus 3 additional prospects in consideration
120%
Quiz Completion Growth
Mobile App Version 2, via AI match-score gating
65%
Engagement Increase
Ambassador Program, after UX redesign
01 — PROBLEM
Why personalization breaks down for smaller beauty brands
Core Problem
The beauty market is cluttered with countless product options;
customers are overwhelmed and can’t get recommendations
tailored to their skin, hair, and lifestyle.
Large platforms like Amazon or Flipkart solve this with vast user data but smaller businesses lack enough data to personalize meaningfully, resulting in a “cold start” problem.
Poor recommendation accuracy from insufficient data leads to
lower conversion, reduced retention, and diminished average order
value (AOV).
Asking users to fill out detailed profile forms for better recommendations caused form fatigue and low engagement. the very data collection needed to solve the cold-start problem was itself getting abandoned.
Who Felt the Pain
B2C
Consumers, ages 15–42+
Overwhelmed by product choice with no tailored guidance
Reluctant to complete long profile forms in exchange for better recommendations
Goal: Get personalized product matches with minimal effort
B2B
Beauty & Wellness Brands
Lacked sufficient data to personalize product recommendations
for their own customers
Needed deep analytics on consumer behavior, product usage,
and purchase patterns
Goal: Match customers with the right products using data-driven confidence
02 — DISCOVERY
01
Proprietary Dataset Analysis
Leveraged a dataset of 200,000+ user recommendations
capturing diverse skin and hair profiles, concerns, and
lifestyle factors.
02
Parameter Analysis
Analyzed key parameters including age, gender, location, skin
sensitivity, and product usage patterns across Indian users.
03
Engagement Research
Identified user pain points related to form fatigue and
low engagement when asked to provide profile data
for recommendations.
04
Expert + User Data Synthesis
Synthesized insights from expert recommendations combined
with actual user product usage for a more reliable, richer
recommendation engine.
KEY INSIGHT
04 — ECOSYSTEM
Five connected surfaces, one recommendation engine
03 — STRATEGY
Solving the cold-start problem by making data collection worth it
01
02
03
05 — SYSTEM
Colors sampled directly from the real B2B Dashboard and Ambassador Program screens, not re-specified after the fact.
Brand Color
#152A5F
Deep Navy — primary dark backgrounds (Ambassador Program)
#4153E9
Brand Indigo — CTAs, active states
#3E4D98
Brand Indigo — data highlights
#FEF464
Signature Yellow — key data bars, accents
Component Inventory
Built only from components present in the shipped screens, then shared across all five surfaces.
Analytics cards
Donut & bar charts
Reward / challenge cards
Progress trackers
Quiz / match-score cards
Product comparison cards
Plugin widget states
06 — PLATFORMS
Needs → Research → Design Approach → Results, for each of the five surfaces.
01
BRAND ANALYTICS
Provide brands deep analytics on consumer behavior, product usage, and purchase patterns
Enable data-driven decision making with user segmentation by demographics, skin/hair profiles, and lifestyle
Visualize expert and user recommendations to identify trends and opportunities
RESEARCH
Analyzed multi-dimensional data for Indian beauty consumers: skin types, common concerns, product preferences
Worked closely with the co- founder to ideate data representation methods that capture user behavior intuitively
DESIGN APPROACH
Created multiple versions of an interactive dashboard with filters (age, gender, location), product shelves, and brand analytics concentrating on top cities, influencers, and purchase channels
Developed visual micro- personas representing user clusters for targeted marketing
Used rapid prototyping (Figma) and iterative client demos for feedback and refinement
RESULTS
Gained 2+ paying B2B clients, with 3 additional prospects considering adoption
Enabled clients to optimize product marketing and inventory decisions based on detailed user insights
02
MOBILE
NEEDS / BUSINESS REQUIREMENTS
Efficiently collect high- quality, personalized user data to support AI recommendations
Overcome user drop-offs caused by tedious form-filling
Incentivize sustained participation through rewards
RESEARCH
Identified that lack of progress tracking and reward visibility hurt engagement in prior designs
Researched gamification and reward mechanisms suited to Indian beauty consumers
DESIGN APPROACH
Completely redesigned the UX/UI in 2 days, introducing progress tracking and reward-claim buttons, gamified unlockable tasks spanning skincare/haircare/makeup, and a profile overview with referral rewards and support features
Delivered a transparent, engaging user journey that motivated consistent participation
RESULTS
Achieved over 2,000 new user engagements and significantly boosted profile completion rates
Earned strong appreciation from the team and end-users for improved usability
Validated the system’s effectiveness at motivating user activity, forming a solid foundation for the AI engine
03
CONSUMER WEB
NEEDS / BUSINESS REQUIREMENTS
Clearly communicate UB’s mission and product value on landing pages
Bridge the gap between the website and app for brand cohesiveness
Enhance user confidence in data collection to improve form completion
RESEARCH
Addressed usability issues like poor color contrast, a missing “About Us” section, and overwhelming promotional banners
Collected user feedback on ease of navigation and trust- building elements
DESIGN APPROACH
Improved accessibility with WCAG-compliant color palettes, reducing eye strain
Replaced sequential banners with user-controlled carousels and clear CTAs
Added transparent explanations of data collection benefits
RESULTS
Increased user trust and data submission rates
Strengthened brand identity through cohesive design language and flows
04
V1 → V2
NEEDS / BUSINESS REQUIREMENTS
Extend UB’s personalized recommendations to mobile users
Incorporate Ambassador Program rewards and data collection into the app itself
Increase profile completion and product engagement through AI match scoring
RESEARCH
Analyzed gaps in user engagement from initial mobile app trials
Explored behavioral triggers for profile quiz completion and product discovery
DESIGN APPROACH
Version 1 — Ported the Ambassador Program design to the app, but engagement stayed limited despite available e-commerce products.
Version 2 — Implemented an AI-powered match-score system that locked product matches until profile completion, incentivizing quiz participation; enhanced product reviews and comparisons to support informed buying.
RESULTS
Version 2 improved quiz completion by 30%, boosting personalization and purchase intent
User interaction with matches and product content increased significantly, enhancing retention and conversion
05
Plugin
FUTURE SAAS
NEEDS / BUSINESS REQUIREMENTS
Offer brands an integrable AI recommendation plugin for
e-commerce platforms
Enable smooth, user-friendly skin profile inputs (face analysis or questionnaires)
RESEARCH
Evaluated form-builder tools (Typeform, Mailform) for seamless data capture
Designed intuitive flows matching product catalogs to user profiles
DESIGN APPROACH
Conceptualized a scalable, AI-driven plugin promising personalized recommendations through minimal user effort
Aligned UI with UX best practices for data collection ease and engagement
RESULTS
Developed a compelling vision and framework for future SaaS offerings within UB’s ecosystem
07 — BRAND & CAMPAIGN
UB’s design work extended past the five core platforms into brand identity, campaigns, and stakeholder communication.
“Consistent branding wasn’t only a product-design problem it had to hold up across social content, sales collateral, and investor conversations too. I extended the same navy/indigo/yellow system and typographic language from the product into every one of these surfaces, so a partner meeting UB through a pitch deck and a user meeting UB through the app encountered the same brand.”
Typography & Color System
The type scale and full color system (grays, native
black/white, primary/secondary brand colors and their tints) underpinning both product UI and marketing materials.
Social Media & Campaign Posts
Instagram campaign posts, logo lockup variations, and
promotional graphics used across UB’s marketing channels.
Blog Post & Story Templates
Reusable templates for blog content and Instagram Stories, keeping content production fast and on-brand.
PDF One-Sheets
Case-study-style PDF sheets used for partner and client- facing communication.
Email Marketing
Templated HTML email layouts for launches and event
announcements, reusing the same brand system as the product.
Pitch Deck & Presentation Design
Investor and stakeholder slides covering problem, platform, traction, and business model — iterated across V1–V3 as the story evolved.
08 — DECISIONS
DECISION 01
Turning Data Collection Into a Reward Loop
CHALLENGE
SOLUTION
WHY THIS MATTERS
DESIGN IMPACT
DECISION 02
Mobile V2
CHALLENGE
SOLUTION
WHY THIS MATTERS
DESIGN IMPACT
DECISION 03
CHALLENGE
SOLUTION
WHY THIS MATTERS
DESIGN IMPACT
Increased trust and data submission rates.
DECISION 04
CHALLENGE
SOLUTION
WHY THIS MATTERS
DESIGN IMPACT
09 — TRADE-OFFS
CONSTRAINT
TRADE-OFF
ACCEPTANCE
Needed to spend design effort on data
collection UX, not just recommendation UI
Engagement design became as important as the AI itself
Users fatigued by long profile forms
Couldn’t just ask for more data — had to
make giving it rewarding
Gamification and micro-incentives over blunt form design
Five products, one brand
Each platform needed its own UI patterns
(dashboard density vs. mobile simplicity)
Shared design system with platform-
specific adaptation
Rapid iteration under real client feedback
Ambassador Program redesigned in just
2 days
Speed was necessary to keep pace with live
user engagement data
10 — OUTCOME
2+
Paying B2B Clients Onboarded
3 more in consideration
40%
Ambassador Program Engagement Increase
After the two-day UX redesign
30%
Mobile App V2 Quiz Completion Growth
Via AI match-score gating
✓
Onboarded 2+ paying B2B clients, with 3 additional prospects considering adoption, powered by 200,000+ user profiles
✓
Increased Ambassador Program engagement by 40%, improving data quality feeding the AI engine
✓
Improved Mobile App (V2) quiz completion by 30%, boosting personalization and purchase intent
✓
Achieved 2,000+ new user engagements through the redesigned Ambassador Program
✓
Improved website form completions and user retention through UI/UX and messaging fixes
✓
Developed a compelling vision and framework for a future Plugin- based SaaS offering
✓
Extended a single brand system across five products plus all marketing, social, and presentation materials
11 — REFLECTION
Small UX Improvements Can Have Outsized Impact
What I’d Do Differently