# Riddhima Sharma, UX Designer

> UX designer with 7 years of design experience; at Deloitte I work across financial services, retail and government. I design agentic and conversational AI products using behavioural design and UX research, and I own the work from concept to delivery. I am also an illustrator. Portfolio: https://riddhima.ca/

## At a glance

- Location: Toronto, Canada. Canadian permanent resident (no sponsorship needed). Open to remote roles
- Current role: UX Designer at Deloitte (2022 to present)
- At Deloitte, also Product Owner and UX Designer for a client reporting platform for investment management (April to July 2026); the platform was sold to a leading investment management firm
- Also: Product Management Mentor at MyCaptain (2025 to present), mentor rating 4.6
- Before: Illustrator and Visual Designer (2019 to 2022); Design Intern at Ogilvy (2019)
- Education: Bachelor's in Design (Communication Design), Banasthali University
- Languages: English, Hindi
- Résumé (PDF, tagged and ATS-readable): https://riddhima.ca/assets/riddhima-sharma-cv.pdf?v=2026-10-12c
- LinkedIn: https://www.linkedin.com/in/riddhsharma
- Email: riddhima.in@gmail.com

## Skills

- Agentic UX and AI product design: agent workflows, human-in-the-loop review, showing where every AI answer came from, conversational AI, prompt engineering
- Behavioural design and UX psychology: reducing cognitive load, designing for trust
- Research: qualitative and quantitative research, persona analysis, usability testing
- Interaction design: user flows, information architecture, wireframes, high-fidelity and interactive prototypes
- Design systems and accessibility (WCAG 2.1)
- Collaboration: engineering and product leadership, stakeholder facilitation, developer handoff, mentoring designers
- Storytelling and illustration: 200+ illustrations for brand campaigns and a bilingual children's book
- Tools: Figma (Figma Make, FigJam, Dev Mode), Gemini, Stitch, Lovable, Claude, ChatGPT, Miro

## Case studies

All three are on https://riddhima.ca/ with a 30-second version and a deep dive. Client and platform names are changed and screens are recreated, to protect confidentiality.

### Personalized Retail (My Wardrobe)

- Role: Lead UX Designer. Timeline: a 2-week pitch.
- Problem: a heritage fashion retailer was losing Gen Z shoppers to stores that feel made for them.
- What I designed: My Wardrobe, a personal corner of the store run by AI agents that learn each shopper's taste. One default agent (Style Buddy) plus agents shoppers create themselves.
- Research: 84% of Gen Z surveyed want a store made for them, 76% trust peers over ads, 92% swipe fluently.
- Result: the concept anchored the pitch that won the retailer as a Deloitte client. A multi-billion-dollar win.

### Agentic Onboarding (Plumbline)

- Role: Lead Product Designer. Timeline: 8 weeks.
- Problem: setting up a workplace retirement plan took up to 12 months, 26 forms and about 275 fields, for answers the company already had in its documents.
- What I designed: onboarding built around those documents. Upload once; agents draft every form and show the source of each answer; people review, correct and sign.
- Design principle: users do not distrust agents, they distrust opacity. Judgment stays with people.
- Result: about 80% less effort to onboard a plan. Won an internal innovation challenge (an 8-figure ARR opportunity). Now being built for one of the largest US retirement providers.

### RFP Agent (Lantern)

An RFP (request for proposal) is the document a company sends out when it wants firms to bid for its work.

- Role: UX and interaction design. Timeline: 4 weeks.
- Problem: proposal leads were pasting RFPs into general chatbots and getting pages of text back, with nothing to check it against and no knowledge of the firm's past bids.
- What I designed: an RFP agent inside the firm's workspace. It starts from the RFP, ties every claim to its source (one click opens the exact line), shows its reasoning before recommending an approach, pressure-tests the draft from six reviewer perspectives, lets people tick any part of the page (a section or a single card) to ask about in chat, with suggested questions that change to match, so the chat already knows which part of the bid someone means, saving the typing and explaining, and leaves a clear blank wherever a person has to decide.
- Process: four versions; each one taught us what to change next.
- Result: strong results in testing with subject-matter experts and future users; now in build.