RecomX

RecomX is your secret weapon for soaring sales; just plug it in, customize it right away, and only pay when we generate income. AI is not required. Only outcomes. Together, let's win.

Comment

Inspiration

Product discovery is broken in the digital economy of today. Businesses struggle to provide the appropriate items to the right people, while customers are overloaded with options. In order to address this, we intended to make personalized product suggestions available to all businesses, regardless of their size or level of technological expertise—just like Amazon, Netflix, or Spotify. RecomX was created with the goal of democratising state-of-the-art recommendation systems with no expense associated with AI technology.

What it does

RecomX is a plug-and-play personalization engine that delivers real-time product recommendations tailored to individual users. It supports:

  • 📦 API-based integration with any e‑commerce or content platform
  • ⚡️ Real-time AWS Personalize inference
  • 🔐 Premium paywall logic to restrict features for non‑subscribed users
  • 📊 Transparent analytics and performance tracking
  • 🔁 Revenue-sharing model: Pay only when our recommendations drive conversions

How we built it

  1. AI Core: AWS Personalize for training & serving user‑item recommendation models
  2. Backend: AWS Lambda + API Gateway with a secure REST API
  3. Frontend: React.js with Chakra UI for a dynamic, mobile‑friendly UX
  4. User Access Control: Premium feature gating via simple ID‑based logic
  5. Design: Clean UI + Unsplash API for enriched product visuals
  6. Deployment: GitHub for code, Vercel for initial frontend, then AWS Lambda for API

Challenges we ran into

AWS Personalize setup: Understanding schema requirements, dataset import formats, and training workflows took time.
Rate limits & API keys: Managing usage quotas for the Unsplash API during image generation.
Vercel issues: Initial deployment failed due to incorrect proxy configs and .env handling.
Git conflicts: Uploading a large codebase and dataset with 3,000+ files via Git CLI had multiple sync and merge issues.

Accomplishments that we're proud of

  • ✅ End-to-end system built in under 4 days, from ideation to deployment
  • ⚙️ Clean separation of frontend, backend, and AI logic
  • 📡 Real-time personalized recommendations with zero latency using AWS
  • 💡 Designed a performance‑aligned business model that’s scalable & monetizable
  • 🎨 Eye‑catching UI with enriched visuals and premium feature gating

What we learned

  • How to work with AWS Personalize, Lambda, and API Gateway at scale
  • The importance of a modular architecture for scalability and maintainability
  • Techniques for debugging real-time deployment issues under pressure
  • The value of good UI/UX in enhancing product experience
  • Git and deployment best practices using GitHub & Vercel under tight deadlines

What's next for RecomX

  1. Dashboard for analytics: Show impact metrics like CTR and conversions
  2. OAuth-based user management and advanced entitlement logic
  3. AutoML retraining pipelines for continuous model improvement
  4. Marketplace SDK so startups can integrate RecomX with one line of code
  5. Web3 smart contract for decentralized, performance‑based revenue sharing
  6. SaaS launch with tiered pricing and open API access

Built With

Share this project:

Updates

Submission history