Isabel Escobar Rivas

I designed and implemented a multi-agent AI platform that transforms fashion inspiration into personalized shopping recommendations. Users provide an image or text prompt, and a Vision Scout agent extracts style attributes while a Style Curator agent assembles coordinated looks across budget, mid-range, and premium price tiers.

The system creates a persistent StyleProfile that captures user preferences and supports increasingly personalized recommendations. I contributed to the product concept, agent workflow, structured data model, technical architecture, API integration, and prototype development. The project was recognized as a 3rd-place finalist in the Complete AI Hackathon hosted by lablab.ai and Complete.dev.

I developed a hackathon prototype that connects transparent creator attribution with instant USDC micropayments on Arc Testnet. The application registers influencers, generates unique campaign links, records click and conversion events, resolves multi-touch attribution, and initiates blockchain payouts through Circle developer-controlled wallets.

I implemented a recency-weighted attribution model that allocates conversion value according to the timing and influence of eligible interactions, while preserving a human-readable rationale for each decision. The application includes an evidence timeline, a payout-status dashboard, REST API endpoints, transaction tracking, and Arcscan links for independent settlement verification.