SheGlow Concierge – AI-Powered Skincare Routine Optimizer
Project Overview
SheGlow Concierge is an AI-powered skincare routine optimizer designed to deliver personalized daily skincare plans to users. The application intelligently analyzes user profiles (skin type, concerns, lifestyle, and goals) and generates tailored product recommendations using generative AI.
Beyond personalization, the app integrates with calendar applications to sync reminders, helping users stay consistent with their routines. By combining AI with serverless cloud infrastructure, SheGlow Concierge demonstrates how advanced beauty tech can enhance self-care while showcasing modern AWS development practices.
Key Features
Personalized Routine Builder: Uses Amazon Bedrock (LLMs) to create daily skincare routines tailored to each user.
Product Recommendations: Matches skincare products to user profiles with explanations for why they fit.
Routine Calendar Sync: Integrates with calendar apps (Google Calendar, Outlook) so routines become part of the user’s daily schedule.
Progress Tracking: Users can log adherence and improvements, stored securely in DynamoDB.
Scalable Architecture: Built on a serverless AWS stack (API Gateway, Lambda, DynamoDB) to ensure performance and cost-efficiency.
Modern DevOps Practices: Deployed with AWS SAM + CloudFormation, version-controlled with GitHub, and automated through CI/CD pipelines.
Frontend: Streamlit app for clean, user-friendly interaction.
Architecture
Frontend: Streamlit web app
Backend: AWS Lambda + API Gateway
Database: DynamoDB (user profiles, logs, product data)
AI Layer: Amazon Bedrock (Generative AI for personalized recommendations)
Infrastructure as Code: CloudFormation templates for consistent deployment
CI/CD: GitHub Actions → AWS CodePipeline → Automated Deployments
Challenges & Solutions
Challenge: Integrating AI recommendations into a structured skincare routine.
Solution: Used Bedrock prompt engineering with DynamoDB-backed data to structure responses.
Challenge: Making the app "sticky" for daily use.
Solution: Added calendar sync integration so users receive timely reminders.
Challenge: Deploying quickly while maintaining best practices.
Solution: Adopted Infrastructure as Code (CloudFormation) and a CI/CD pipeline for reliable deployments.
Impact
The SheGlow Concierge project demonstrates the intersection of AI, cloud engineering, and real-world user needs. It not only highlights my technical expertise in AWS Developer tools, DevOps practices, and AI integration, but also showcases my ability to design solutions that improve people’s lives.
Tech Stack
Cloud: AWS (Lambda, API Gateway, DynamoDB, Bedrock, CloudFormation)
DevOps: GitHub Actions, AWS CodePipeline
Frontend: Streamlit
Languages: Python, YAML (IaC)
Repo
