AI built for real campus life, not screen time
Uplifty is an AI-powered social mobile platform for college students, available on iOS and Android. Our system uses machine learning and behavioral science to help students build genuine friendships, discover meaningful challenges, and grow inside their campus communities—without the toxic engagement loops that define mainstream social media.
Platform Status
Uplifty is in active development with app store submissions to Apple App Store and Google Play Store underway. The core mobile application, backend services, and AI infrastructure are operational.
We are an active Google Cloud enterprise customer with Firebase, Cloud Run, Vertex AI, Cloud Pub/Sub, BigQuery, and Cloud Storage services currently provisioned. Our AI/ML models are in the process of being uploaded and deployed to Vertex AI. We are applying to the Google Cloud AI Startup Program to access the cloud credits needed to run our integrated ML models and API services at the scale required for a multi-campus launch.
Our AI/ML engineering team holds graduate degrees in machine learning, computer science, and data science, with hands-on experience building recommendation systems, NLP pipelines, and mobile-integrated ML services. The broader team spans app engineering, product, design, and go-to-market across more than eighty contributors.
Meet the full team →
Market opportunity
Uplifty’s initial focus is the U.S. higher education market—more than 6,000 colleges and universities serving approximately 19.2 million enrolled students across undergraduate and graduate programs, according to the National Student Clearinghouse Research Center. Every one of those campuses is a self-contained social environment with shared physical spaces, high social need, and strong peer network effects—precisely the conditions where a campus-first social platform scales fastest.
Press coverage
Uplifty has received international media coverage ahead of its public launch. A major international English-language publication with a global readership covered the platform’s mission and early traction, describing Uplifty as a “post-social media cultural response” designed to help students build genuine community offline rather than exploit their attention online. The coverage highlighted the platform’s encrypted architecture, privacy-first design, and founder Scott Amyx’s vision for a fundamental reset of how social technology serves young people.
The Platform
Uplifty is not a web application.
It is a native mobile experience—built in React Native and Expo—where every feature is designed around the physical reality of campus life.
Students meet through real-world challenges.
They track friendships through Friend Chain.
They collect moments in Friendship Passports.
They discover what is happening across their campus through an AI-curated feed.
The AI layer runs entirely behind the scenes. Students do not interact with a chatbot or an AI assistant. Instead, the system learns from how students engage—which challenges they complete, which friends they connect with, what content they stop on—and uses those signals to make the experience progressively more relevant and valuable over time.
Three Proprietary Intelligence Systems
Uplifty’s AI platform is composed of three independent but interconnected intelligence systems. Each handles a distinct layer of the personalization problem.
Human Growth Intelligence
Builds a dynamic model of each student’s growth journey using behavioral sequence modeling to track engagement patterns, interaction history, and progress trajectories over time. Identifies moments where a student would benefit from a new challenge or connection.
Personalized AI
Generates real-time recommendations and challenge suggestions using collaborative filtering combined with live behavioral context. Adapts continuously to support habit formation and sustained engagement with the things that actually matter to each student.
Content Intelligence
Ranks and surfaces content using embedding-based semantic similarity to match material to each student’s goals, interests, and behavioral state. Filters low-value material and ensures the campus feed is consistently constructive, safe, and relevant.
How the System Works
User actions on the iOS and Android apps generate behavioral events that flow in real time into the intelligence pipeline. Cloud Run handles API routing and authentication at low latency. Cloud Pub/Sub streams behavioral events to the intelligence layer without blocking the client experience. Firebase manages authentication, real-time data sync, and push notifications via Cloud Messaging.
The three AI systems process behavioral signals and return personalized decisions—which challenge to surface, which friend to suggest, how to rank the campus feed—within each app session. Vertex AI powers model training, real-time inference, and continuous evaluation. BigQuery enables long-term behavioral analysis and model improvement at scale. Google Maps SDK supports location-aware campus features including proximity-based challenge discovery.
The system separates real-time decisioning from long-term model learning, ensuring responsiveness for students while allowing models to improve continuously without disrupting live service.
Cold-Start Strategy
New platforms face a fundamental challenge: recommendation models require behavioral data that does not exist at launch. Uplifty addresses this directly.
Pre-launch model training uses public domain behavioral and wellness datasets—including the GoWellness dataset—to establish baseline recommendation quality from day one. Challenge recommendation and content ranking models are initialized using these datasets and refined through challenge category matching and campus context signals available at onboarding. As students engage with the platform, live behavioral signals progressively replace these seeds, continuously improving recommendation quality as the user base grows.
Built on Google Cloud
Uplifty’s production system runs entirely on Google Cloud Platform. The decision to build on GCP was not incidental—the combination of Firebase’s mobile infrastructure, Vertex AI’s ML tooling, and Google’s content moderation and mapping APIs creates an integrated stack that no other cloud provider offers with the same depth for a mobile-first AI product.
Our CI/CD pipeline runs through GitHub Actions with Firebase App Distribution handling APK delivery, giving the engineering team a continuous deployment workflow fully integrated with GCP services from development through production.
Firebase
Authentication, Cloud Messaging for push notifications, Firestore for real-time data sync, and App Distribution for CI/CD delivery across iOS and Android.
Cloud Run
Serverless, auto-scaling API layer serving all mobile client requests with low-latency response times.
Cloud Pub/Sub
Event streaming pipeline that decouples behavioral signal ingestion from real-time AI inference without blocking the client.
Vertex AI
Hosts model training and real-time inference serving for HGI, PAI, and Content AI. Drift monitoring pipelines evaluate model performance continuously in production.
BigQuery
Behavioral analytics warehouse enabling long-term model improvement, product insights, and well-being outcome measurement at scale.
Cloud Storage
Media storage for user-generated content, challenge assets, and Friendship Passport stamps.
Cloud Content Moderation API
Google’s content moderation APIs provide a runtime safety layer on user-generated content and AI-generated recommendations before they reach students.
Google Maps SDK
Powers location-aware campus features including proximity-based challenge discovery and the Campus Map. Currently in integration.
A New Kind of Advertising
The standard social media business model optimizes for time-on-screen.
More time means more ads.
More ads means the algorithm is rewarded for surfaces that generate anxiety, comparison, and compulsive scrolling.
Uplifty rejects this entirely.
Our advertising model is built on well-being signals: challenge completions, genuine social connections formed, communities joined, and personal goals met. Advertisers reach college students through placements that appear alongside positive, constructive activity—not algorithmically amplified distress.
This means our AI is never incentivized to maximize unhealthy engagement.
The model that drives the business is the same model that improves user outcomes.
Uplifty’s ad inventory is categorically brand-safe and ethically differentiated from every major competitor in the college demographic.
“We measure success by what students accomplish, not by how long we keep them looking at their phone.”
Ethical and Responsible AI
Responsible AI is not a policy document at Uplifty.
It is a runtime constraint.
The same systems that power personalization enforce behavioral guardrails, safety filtering, and bias-aware ranking at every inference call. Google Cloud’s Content Moderation APIs provide an additional safety layer at the infrastructure level, screening user-generated content and AI outputs before they reach students.
Content safety filtering
Google Cloud Content Moderation APIs screen all user-generated and AI-generated content before it reaches students, with additional application-layer filters on AI recommendations.
Bias-aware ranking
Personalization models are evaluated for demographic bias, with adjustments applied in the ranking layer before recommendations are served.
Behavioral guardrails
The system is designed to surface growth-oriented content and connections, with hard constraints preventing outputs that would promote comparison, dependency, or harmful behavior patterns.
Human feedback loops
Students can flag, dismiss, or provide direct feedback on AI recommendations. These signals flow back into model evaluation and retraining pipelines.
Privacy-first design
Behavioral modeling is performed on anonymized signal data. Uplifty does not sell user data to third parties.
Continuous monitoring
Model outputs are monitored in production for drift, quality degradation, and alignment with well-being metrics using Vertex AI’s built-in evaluation pipelines.
Built for Impact, Not Engagement
Uplifty exists to prove that social technology can make people’s lives measurably better.
By combining behavioral science, machine learning, and a principled approach to monetization, we are building a new category of social platform—one where the AI serves the student, not the advertiser’s click-through rate.
Our mission is to help college students build the real friendships, habits, and communities that will define who they become—and to do it at scale, on the phones already in their pockets.
This is not a different feature set.
It is a different philosophy.