Andre Espinoza

Built with OpenAI GPT, Python, Flask, and Figma for the front-end prototype. My responsibilities included user research, API configuration, and training the model. Two rounds of user testing showed students could find courses in under 5 minutes. The biggest challenge was ensuring data-specific responses — solved by building 500+ unique training queries and switching from Google Cloud to the updated OpenAI API platform on short notice.

Developed a conceptual coding scheme using AI to categorize reviews into Positive, Negative, and Neutral feedback types with sub-codes (Compliments, Complaints, Recommendations, etc.). Findings: both restaurants show ~36% positive compliments — Lucky Robot had fewer service complaints, Cabo Bob's had more actionable constructive criticism. Delivered data-backed recommendations for staff training and feedback utilization.

Surveyed 64 young adults at South Congress and UT Austin. Key findings: young adults are motivated by networking, personal fulfillment, leadership opportunities, and civic engagement. Proposed an influencer-driven public volunteer event model, recurring leadership programs, and social media-first outreach strategy to drive sign-ups and long-term retention.

5-month collaborative project using Figma, Miro, wonder tables, and mind maps to understand climate injustice as a design challenge. Final design included: explore page, search, org list, org profile, and user profile — each built for engagement and ease. The main challenge was adapting the dating-app interaction model ethically to the context of charity without losing the addictive usability.

Research included interviews with locals and patrons, analysis of historical documents and photos, and study of Austin urban development trends. Built in Unity and Blender. Final experience lets users explore the venue across different eras, interact with historical artifacts, and engage with live audio-visual elements. User testing focused on emotional authenticity — does this feel like the real Broken Spoke?

Process moved from app audit → pain point mapping → wireframing for both mobile and Smart TV → high-fidelity Figma prototypes. Core improvements: cleaner navigation hierarchy, improved search UX, and consistent cross-platform branding. The challenge was enhancing usability without disrupting brand recognition — users needed to feel the redesign was still unmistakably Audi.

Design challenge: make a dense, multi-device system feel simple and intuitive. Process: ecosystem audit → integration opportunity mapping → interface design across home screen, room controls, camera feeds, thermostat, and music. Clear visual hierarchy and logical layout allowed users to manage complex systems with minimal cognitive load.