EUJAY FLOW

Inspiration

What it does

Inspiration

As Decentralized Physical Infrastructure Networks (DePIN) grow globally, managing edge node hardware under fluctuating network latency, hardware overheating, and bandwidth bottlenecks becomes a critical challenge. Eujay Flow was built to provide an autonomous, AI-driven orchestration layer that ingests physical telemetry, leverages cloud intelligence for decision-making, and maintains network resilience even during offline or weak-network conditions.

What it does

Eujay Flow is an enterprise DePIN orchestration platform submitted under Track 5: EdgeAgent.

Perceive via Edge Sensors: Collects live edge telemetry (bandwidth usage, system load, network latency, and node temperature) from connected hardware.
Reason via Cloud APIs: Passes edge metrics to Alibaba Cloud's DashScope API utilizing the Qwen-Turbo LLM model to analyze node health and evaluate optimal routing strategies.
Act Locally: Receives structured JSON decision commands from Qwen AI to automatically adjust edge node behavior (e.g., bandwidth throttling, load re-routing, or scheduled cache refreshes).
Graceful Degradation: Features an automated local fallback strategy (APPLY_LOCAL_EDGE_CACHE_POLICY) when API timeouts or network disconnects occur.

How we built it

Frontend: Built with React, Vite, and Tailwind CSS, hosted on Vercel (https://eujay-flow.vercel.app/).
Backend: Built with Node.js and Express, running on an Alibaba Cloud ECS instance managed by PM2.
Database: PostgreSQL managed via Prisma ORM on Alibaba Cloud.
AI Engine: Integrated with Alibaba Qwen Cloud (DashScope API) for model reasoning over telemetry data.

Challenges we overcame

Deploying a hybrid edge-cloud setup required strict error handling between HTTPS client environments and backend nodes. Designing a seamless "graceful degradation" fallback ensured that if the cloud AI link drops, local edge nodes continue operating without crashing the system.

Accomplishments that we're proud of

Fully deploying a full-stack DePIN platform on Alibaba Cloud ECS with a live PostgreSQL database.
Successfully connecting Qwen Cloud AI via DashScope API to process physical telemetry and generate JSON action payloads.
Building a production-ready, dark-mode dashboard with interactive AI diagnostics and weak-network resilience.

What's next for Eujay Flow

Expanding on-chain smart contract integration for automated token rewards based on AI-verified node uptime.
Supporting multi-region Alibaba Cloud deployments for localized AI inference closer to physical edge devices. ## How we built it