Personal Projects / 2025
AI Persona Generator
Designed and built an AI tool that turns raw research transcripts into personas and opportunity solution trees — synthesis in minutes instead of weeks. Built end-to-end with Cursor and Gemini.
Design Highlights
Overview
AI Persona Generator is a tool I vibe-coded using Cursor, powered by Google Gemini, that bridges the gap between raw research insights, opportunity mapping, and design exploration. The tool automatically generates detailed personas and opportunity solution trees from research transcripts and insights. I built it end-to-end as a designer, from concept to deployed product, using Cursor's AI-assisted development environment.
Challenge
Product teams faced significant challenges in manually synthesizing research findings into actionable personas and opportunity solution trees. The process typically took weeks of manual effort, often resulting in inconsistent outputs and missed insights. This bottleneck delayed design exploration and strategic decision-making.
Approach
I vibe-coded the entire application using Cursor, going from concept to deployed product in days rather than weeks. I used Google Gemini to analyze research transcripts and extract patterns, user needs, pain points, and opportunities, then designed an interface for uploading source data and refining the generated outputs.
Solution
Created a web application that accepts research transcripts and insights as input, processes them through Gemini AI, and generates detailed personas — demographics, goals, pain points, and behavioral patterns. The system also automatically constructs opportunity solution trees, mapping user needs to potential solutions and design opportunities.
Impact
Reduced research synthesis time from weeks to minutes (99%+ time reduction). Cut opportunity mapping from hours to minutes. Enabled teams to explore more design directions by removing the research synthesis bottleneck. Generated higher quality and more detailed insights by systematically analyzing all research data rather than relying on manual synthesis.
Key Learnings
I learned through vibe-coding with Cursor that designers can ship real products without a traditional development workflow. AI can synthesize research systematically when it is properly guided, but human review remains essential. Structured outputs such as personas and opportunity trees helped turn the synthesis into action.