Designing stuff marks my existence

github

Bhrigu Kumar Deka  • Peerlist

<p>DimLight — AI-Powered Sleep Monitoring &amp; Coaching Platform MERN Stack | Deployed Full-Stack Application Tech Stack Frontend: React, Vite, TailwindCSS, Recharts Backend: Node.js, Express.js, MongoDB (Mongoose) Auth &amp; Security: JWT, bcrypt, email verification &amp; password reset Deployment: Vercel (frontend), Render (backend), MongoDB Atlas Problem Sleep data is fragmented, unstructured, and hard to interpret for users Solution Built a centralized platform to log sleep, analyze patterns, and deliver personalized insights Key Features Sleep logging with quality, mood, stress, and lifestyle factors Multi-factor sleep scoring algorithm (duration, quality, consistency, lifestyle) Rule-based pattern detection for sleep issues (stress, caffeine, inconsistency) AI sleep coach using Google Gemini with caching &amp; rate limiting Google Fit integration via OAuth for external sleep data sync</p>

DimLight

Tools Figma, Miro, Balsamiq, Pen & Paper Figma Plugins: Iconify, Automatic Style Guides, Mockup Plugin The Problem Modern urban shoppers are overwhelmed. Users told me, "I open four apps just to check if my orders are on track—it's so tiring". This fragmented experience leads to anxiety, disorganization, and financial loss when users miss return deadlines because they can't remember where they bought an item. The Solution A unified, intelligent app that centralizes all online orders, return timelines, and outfit suggestions. It eliminates the need to depend on different sites and messages by combining disconnected touchpoints into one centralized platform. Highlights Unified Tracking: A central hub to track purchases across platforms like Myntra, Amazon, and Flipkart. Return Window Tracker: Features a countdown badge (e.g., "Return ends in 2 days") to ensure users never miss a refund. Smart Suggestions: A "Match & Style" AI summary that helps users visualize how new items fit their existing wardrobe. Calm UI: Designed with a dark blue palette (#030719) to evoke trust, responsibility, and calmness. The Result The high-fidelity prototype was tested with 6 users and achieved a System Usability Scale (SUS) score of 97.5. Feedback highlighted that the system was "Excellent," "Easy to use," and provided a "Good, convenient system".