Bridge

• 

  web demo

Inspiration

54 million older adults in America are lonely. 45 million college students lack mentorship. That's 99 million people searching for connection but missing each other completely. The loneliness epidemic costs U.S. healthcare $6.7 billion annually. It increases mortality risk by 26%. It's as deadly as smoking 15 cigarettes a day. Meanwhile, 73% of college students report feeling overwhelmed, anxious, and directionless, desperately needing guidance that their overworked professors and stressed-out peers can't provide. We saw two crises that were actually one solution. Older adults have wisdom, time, and a need to feel valued. Students have energy, curiosity, and a hunger for mentorship. The problem wasn't that these groups didn't want each other, it's that they had no way to find each other. Dating apps proved that AI-powered matching could connect strangers into meaningful relationships. Social media proved that digital platforms could facilitate real-world community. We asked: why doesn't this exist for mentorship? Bridge was born from that question. We're not building another social network. We're building social infrastructure, a way to systematically reconnect generations that society has systematically separated.

What it does

Bridge is an AI-powered mentorship platform that connects older adults and college students for genuine, intergenerational relationships. It’s not about likes, followers, or swipes—it’s about shared humanity. Older adults, our mentors, create profiles that highlight their experiences, passions, and what they’d love to share. Students create their own, sharing their goals, aspirations, and the guidance they seek. Then our matching algorithm goes to work—analyzing interests (40%), personality compatibility (30%), motivation alignment (20%), and location (10%)—to suggest pairings that actually make sense. When a strong match appears, a conversation can begin instantly. Students can type messages, but older adults often prefer speaking to typing—so we built something special: a live voice-to-text feature that lets mentors simply talk, and the app transcribes their words in real time. It feels natural, human, and effortless—like a phone call that slowly becomes a friendship.

How we built it

Frontend: Next.js 16 + TypeScript for type-safe, scalable, server-rendered applications Tailwind CSS with React Bits for responsive, accessible UI that works equally well for seniors and students Lucide Icons for clean, intuitive visual design

Backend: Firebase Authentication for secure user management Cloud Firestore for real-time database synchronization Firebase Storage for profile photos and audio data with automatic CDN optimization Real-time listeners to support instant messaging and updates

Matching Algorithm: The compatibility model runs server-side and recalculates dynamically as users update their profiles. It uses a combination of interest overlap (Jaccard similarity), personality vector distance, and goal complementarity to produce a single unified compatibility score.

Voice-to-Text System: To improve accessibility for older adults, we implemented real-time voice input that converts speech to text instantly within the chat interface. This allows users to communicate naturally without relying on typing. The system uploads short audio segments to Firebase Storage, streams them for transcription, and returns results live—creating a smooth conversation flow across generations.

Challenges we ran into

1. Accessibility and Tech Literacy Designing for both 70-year-olds and 20-year-olds was the hardest problem. Many older users found typing difficult or stressful. To solve this, we integrated voice-to-text and voice message recording features. We also optimized the interface with larger touch targets, readable text, simplified navigation, and clear feedback for every action.

2. Real-Time Messaging Architecture Building low-latency, error-free chat required careful management of Firebase listeners, offline queues, and race conditions. We developed a conversation-based data model with atomic writes, optimistic updates, and auto-cleaned subscriptions to prevent duplication and crashes.

3. Matching Accuracy Early versions of the matching algorithm were too simple (leading to irrelevant pairings) or too complex (causing performance delays). Through testing with over 20 user profiles, we refined the final weighting to balance speed and precision.

4. Building Trust Intergenerational matching requires high trust. We implemented email verification, optional ID verification, content moderation, and a transparent reporting system. Clear UI messaging and simple privacy controls helped users feel safe and confident.

5. Media Handling Profile photo and voice upload performance was another challenge. We optimized image resizing, implemented upload progress indicators, and used atomic updates to maintain consistent user data across Firestore and Storage.

Accomplishments that we're proud of

Created a fully functional, real-time AI mentorship platform accessible across age groups. Developed a voice-to-text system enabling older adults to communicate naturally without typing. Achieved smooth, scalable chat performance with zero manual refresh or delay. Designed a clean, inclusive UI that feels both modern for students and intuitive for seniors. Built a robust matching engine that produces high-quality pairings using behavioral and personality data.

What we learned

Accessibility drives adoption: older adults aren’t afraid of technology—they just need interfaces designed around comfort, not assumptions. Real-time systems are complex: every message involves synchronization, latency, and memory considerations. Trust is a design element: users feel safer when feedback is immediate and language is friendly. Impact and usability must co-exist: social good and product design reinforce each other when built from user needs.

What's next for Bridge

Next 90 Days: Launch a beta pilot at UC Berkeley, pairing 50 students with 50 local older adults. Collect metrics on match quality, engagement time, and loneliness reduction. Add video calls and improve mobile accessibility.

Next 6 Months: Integrate a housing connection feature, allowing verified mentors to offer affordable housing to students. Enable group mentorship for workshops and Q&A sessions. Partner with universities and senior centers to scale adoption. Expand safety infrastructure with AI-based content moderation and real-time support.

Long-Term Vision: Bridge will become the leading platform for intergenerational mentorship—reducing loneliness by 40% among users, helping 100,000 students find mentors, and redefining how technology connects generations.

Built With

• cloud-functions)
• elasticsearch
• fetch.ai-api
• firebase-(authentication
• firebase-admin-sdk
• firestore-database
• git
• next.js-16
• npm
• react-19
• storage
• tailwind-css
• typescript