Product designer · Developer background · AI-assisted
Designing at the speed of AI. Grounded in product thinking.
My software-development background shapes how I approach product design. I use AI to turn complex requirements into structured workflows, translate established design systems into working interfaces, explore ideas through code, and validate experiences before they move further into development.
- AI-assisted product design
- Figma MCP & design-system grounding
- Rapid UI prototyping
- Vibe coding
- Workflow automation & validation
02The end-to-end workflow
Understand, structure, design, prototype, validate, refine.
AI runs through every stage — from an ambiguous brief to a reviewable experience, faster.
- 01
Understand
Requirements, user needs and constraints — pulled from briefs and discussions before any screen exists.
- 02
Structure
Turn those into user journeys, flows, states and decision points — the behaviour, not just the layout.
- 03
Design
Ground the interface in the existing design system — real components, tokens and conventions.
- 04
Prototype
Translate intent into a clickable, coded prototype so behaviour is testable, not just visible.
- 05
Validate
Inspect interactions, states and responsiveness; surface issues and edge cases early.
- 06
Refine
Review with stakeholders, capture the decisions, and iterate toward the right implementation.
03Figma MCP + design systems
From design system to working prototype.
I connect design context with AI-assisted coding so I can explore product interfaces without starting from a blank page every time. The point isn't to generate arbitrary components — it's to ground generated UI in an established design language and review it against real requirements.
Begin from the relevant Figma file — existing components, design tokens, layouts and product conventions.
AI-generated interfaces become genuinely useful when they're grounded in an existing design system and reviewed against real product requirements.
04From conversations to requirements
Turning messy discussions into a brief I can design from.
Before any interface exists, there's a pile of discussion — stakeholder calls, design reviews, scattered notes. I use AI to turn that into a structured brief I can actually design from, then review it against the source rather than trusting the output.
- “…so it needs to work on mobile too, obviously.”
- “Can we reuse the existing filter component?”
- “What happens if the search returns nothing?”
- “Legal still needs to confirm the data we store.”
- action: Priya to check the catalog API limits
- User needFind a nearby store without typing an address.
- FunctionalGeolocation with ZIP/city fallback; results within a radius.
- StatesEmpty, loading, permission-denied, no-results.
- ConstraintReuse the existing filter + list components.
- Open questionConfirm stored-location data with legal.
05AI inside real, shipped products
Not a demo reel — AI experiences I designed and shipped at Cyncly.
Natural-language discovery & catalog filtering
Let people describe what they want in words, and ground the results in the real catalog so suggestions are actually selectable.
AI-assisted configuration & Auto Design
Compress the first pass: let AI generate catalog-valid layouts from a sketch or a few choices, then hand off cleanly into the Pro tools.
Inspiration, refinement & product selection
Design the path from inspiration to selection — refine by conversation, keep choices grounded in what's real and available.
From a design to a real store & booked consultation
Bridge the on-screen design to a physical showroom — find a store, compare it, and book a consultation, with every location state handled.
06The toolkit
Grouped by the job it does.
AI reasoning & coding
GitHub Copilot
AI coding assistance · In-editor scaffolding, refactors and reusable Markdown AI skills.
Claude Code
Agentic AI coding · Builds prototypes and ships real interfaces — this portfolio included.
ChatGPT / GPT
Reasoning · research · docs · Requirements, journeys, PRDs, prompts and copy.
Gemini
Multimodal reasoning · Visual and technical exploration across text and images.
AI-assisted prototyping
Antigravity
AI-assisted prototyping · Rapid, interactive frontend prototypes for early reviews.
Google Stitch
AI UI generation · Alternative interfaces and faster first concepts.
VS Code + Git
Code-based prototyping · Supporting environment for moving toward working interfaces.
Design & communication
Figma / Figma MCP
Design source · prompt-to-prototype · Design work, and the design-system context I feed to AI-assisted coding via MCP.
07The judgment layer
What AI does — and what stays mine.
AI handles
- Generate alternatives
- Draft interface structures
- Suggest states & edge cases
- Produce initial code
- Organize documentation
I own
- User intent
- Product context
- Business priorities
- Technical feasibility
- Accessibility
- Design consistency
- Quality & reliability
AI expands my capabilities. Product judgment gives them direction.
I use AI to reduce repetitive effort, explore more possibilities, and bring ideas into reviewable form earlier. I combine that speed with product thinking, design-system discipline, and hands-on implementation.