RECO.
Beating rehab burnout.
ACL recovery isn't just a physical battle; it's a mental one. We built an AI companion in 48 hours to keep athletes motivated during the most isolating phases of physical therapy.
How we shipped in 48 hours.
Empathy is Not Hypothetical
One of our teammates actively recovering from an ACL tear gave us immediate, zero-BS research. Bypassing hypothetical pain points, we designed strictly for her daily frustrations with isolation and repetitive exercises.
Leveraging Vibe Coding
We embraced Vibe Coding. Offloading boilerplate syntax to AI allowed us to spend 100% of our cognitive bandwidth on technical constraints, product strategy, and prompt engineering.
Flat Structure & Rapid Alignment
We didn't know each other and it was our first time working together. We established a flat structure to move from concept to shipped iOS prototype within an aggressive cutoff, completely eliminating conceptual debates.
Depression kills exercise adherence.
Rehabilitation is an isolating psychological battle. Data shows that 41.8% of athletes experienced anxiety and depression, which are negatively related to home exercise adherence. We needed an intervention system grounded in Self-Determination Theory (SDT).
"The biggest challenge after ACL surgery isn't just physical — it's psychological."
Caumeil B, et al. Reinjury Anxiety and Return to Sport After ACL Reconstruction: Cluster Analysis and Prospective Study Among 162 Athletes.
Autonomy, Competence, Relatedness — the pillars of motivation.
RECO isn't a passive tracker. It's an active intervention system grounded in Self-Determination Theory (SDT). It restores motivation by giving athletes a sense of control over their recovery, clear progress signals, and an emotionally supportive companion.
VITA AI Companion
AI companion trained on PubMed research. It provides evidence-based rehabilitation guidance with expertise in nutrition and recovery science.
Smartphone IMU
Uses phone sensors to measure the angle of leg lift, speed, and amplitude — giving accurate, quantitative feedback on exercise form and progress.
iOS Widgets
Home screen widgets showing recovery day count and daily check-ins, increasing utilization and giving athletes a stronger sense of companionship.
BoKnee — the companion in action.
High-fidelity screens reflecting our core design principle: The UI must feel warm, approachable, and completely devoid of clinical anxiety.
API Integration, UX, & Frontend Development
Owned the technical pipeline, visual language, and frontend execution. I manually engineered prompt constraints for medical viability and defined the entire user journey in Figma.
Constraining the LLM
Action: Integrated GPT-4o-mini API and designed strict system prompts.
Result: Locked VITA into rehabilitation/nutrition domain, eliminating medical hallucination risks.
PubMed Medical Pipeline
Action: Hooked the backend to the NCBI Entrez PubMed API.
Result: VITA pulls advice directly from peer-reviewed papers instead of generic internet data, building immediate user trust.
Protecting Credentials
Action: Deployed Supabase Server-side Key Encryption.
Result: OpenAI and PubMed keys are encrypted and never exposed to the frontend bundle—basic security hygiene.
Frontend & Interaction
Action: Developed key screens (chat, visualization widgets) in Swift/React Native.
Result: Delivered smooth functional components within the 48-hour cutoff.
Visual Design & Empathy (Figma Role)
Action: Defined user flows and iterated visual language entirely within Figma.
Result: Crafted a calm, approachable visual style that addressed the emotional needs of athletes in pain, reinforcing companionship.
What we built with.
| AI MODEL | ChatGPT-4o-mini | OpenAI API |
| MEDICAL DATA | NCBI Entrez PubMed Database | MED-API |
| SECURITY | Supabase Server-side Key Storage | Encryption |
| PLATFORM | iOS Mobile Application | Swift / React Native |
| DESIGN TOOL | Figma (Role: User Flow & Visual System) | Figma |
| FRAMEWORK | Self-Determination Theory (SDT) | Psychology |
"A hackathon stops you from over-designing and starts you building. It proved that empathy, combined with the right AI, can solve real problems, fast."