Juan Borges Vega
EasyMind VR
Independent · Android · VR · Unity · AI · 2024 – Present
A mobile VR mental wellbeing application independently designed, built, and published on Google Play. Integrates speech emotion recognition via a machine learning inference API and a reinforcement learning-based recommendation system. Validated through structured user testing and reviewed by a collaborating psychiatrist.
HIGHLIGHTS
Speech Emotion Recognition & Adaptive Recommendations
Integrated a RESTful API connecting to a Hugging Face inference model for real-time speech emotion recognition, allowing the application to detect the user's emotional state from voice input. Built a reinforcement learning-based recommendation system on top of this, using detected emotional states to adaptively suggest personalized relaxation activities. The combination creates a feedback loop that adjusts recommendations over time based on observed user responses.
Structured User Testing & Psychiatric Validation
Conducted structured user testing with 20 participants, measuring and observing emotional trend indicators across sessions. Results were reviewed and validated by a collaborating psychiatrist, confirming the application's potential as a personal emotional wellbeing tool. The project was awarded 2nd place at the VIII National Undergraduate Thesis Competition by the Venezuelan Computer Society, and received the highest academic grade.
End-to-End Independent Delivery
Managed the complete development lifecycle independently — from concept and architecture through implementation, localization, user testing, and publication on Google Play. Built within a scalable Unity architecture designed to support future feature expansion.
ADDITIONAL CONTRIBUTIONS
- Designed and built immersive VR environments optimized for mobile hardware constraints.
- Implemented activity tracking dashboards displaying emotional trend data over time
- Built breathing activity mechanics within tranquil nature environments as core wellbeing interactions
- Architected the full application within a scalable Unity structure designed for future feature expansion
