React & Node.js Developer Portfolio - JavaScript Projects

educationELLy is a full-stack web application for managing English Language Learner (ELL) students in educational settings. It provides a collaborative platform for ELL and mainstream teachers to track student information, English proficiency levels, and academic progress. Built with React/Redux frontend and Node.js/Express/MongoDB backend, it features secure JWT authentication, comprehensive student profile management (demographics, language details, academic status), and a responsive interface with real-time data access for improved teacher collaboration and student support. Deployed on Kubernetes with containerized microservices and MongoDB for persistent document storage.

Code Talk is a real-time collaborative code editor and messaging platform that enables developers to write code together while communicating seamlessly. Built with React and GraphQL, it features live collaborative editing where multiple users can work on the same code simultaneously, integrated instant messaging with room-based organization, and secure JWT authentication. The app leverages WebSocket subscriptions for real-time updates, Redis pub/sub for scalability, and includes performance optimizations like virtualized lists and code splitting. Deployed on Kubernetes with horizontally scaled server replicas, PostgreSQL for persistent storage, and Redis for real-time pub/sub messaging. Perfect for remote pair programming, code reviews, or team collaboration sessions.

educationELLy GraphQL is an education management system for English Language Learner (ELL) students. It features a React frontend with Apollo Client for GraphQL integration and a Node.js backend with Apollo Server, Express, and MongoDB. The app provides user authentication, full CRUD operations for managing ELL student records (including personal info, educational details, native language, proficiency levels), and role-based access control for teachers and administrators to track and manage their ELL student population. Deployed on Kubernetes with containerized Apollo Server and MongoDB for persistent storage. The educationELLy REST version and GraphQL version use separate MongoDB StatefulSets, allowing independent scaling and data isolation. The GraphQL version benefits from Apollo Server built-in features like query batching and caching.

FireBook is an intelligent bookmark management platform built with React and Firebase that automatically enriches saved links with metadata, screenshots, and AI-generated tags. The app features Google OAuth authentication, real-time Firestore synchronization, and Firebase Cloud Functions that capture webpage screenshots using Puppeteer, extract metadata (titles, descriptions, favicons), and generate contextual tags via Google Natural Language API. Users can search bookmarks with Algolia instant search, organize collections with collaborative sharing (viewer/editor permissions), edit bookmarks in-app, and import/export data as JSON. Built with Vite, Tailwind CSS, and deployed on both Firebase Hosting and Kubernetes, FireBook demonstrates serverless architecture, cloud automation, and modern React development patterns.

Bookmarked is an AI-powered bookmark manager built with React Hooks and LangChain that automatically organizes your saved links with intelligent tagging. Using Llama 3.2 3B Instruct (with GPT-4o-mini as fallback), the app analyzes bookmark titles, URLs, and descriptions to generate 3-5 relevant tags instantly—either individually or in batches. Users can save, rate, and favorite bookmarks while leveraging full-text search and tag-based filtering to find what they need. Built with React Context API and useReducer for state management, the app connects to a Neon PostgreSQL database with a normalized tags schema and PostgreSQL-backed caching to minimize API costs. Features a cost-efficient hybrid AI architecture with Llama 3.2 3B Instruct (Q4_K_M quantization) via llama.cpp for local inference, falling back to OpenAI when needed. Deployed on both Vercel serverless functions and Kubernetes with containerized microservices and a shared AI gateway for high-performance production inference.

IntervalAI is a neural-enhanced spaced repetition learning system that uses machine learning to optimize vocabulary review timing for maximum retention. The app combines the proven SM-2 algorithm with a 10-layer deep neural network (18,305 parameters) using 51 engineered features including forgetting curve modeling, interaction features, and cyclical time encoding. The advanced model achieves 96.1% improvement over baseline methods (MAE: 0.0735 vs 1.8889 days). Built with React/Redux frontend and Node.js/Express/MongoDB backend, it features WebGPU-accelerated ML inference (<1ms) on the client, real-time predictions of optimal review intervals, TensorFlow.js for machine learning, and built-in A/B testing to compare algorithm performance. Deployed with NVIDIA Triton Inference Server on Kubernetes for high-performance production inference. Perfect for language learners seeking personalized, scientifically-optimized study schedules.

Pop!_Portfolio (Portfolio Orchestration Platform) is a Kubernetes-based orchestration platform that manages and monitors all portfolio applications as containerized workloads. Built with React and Node.js, it provides real-time pod status tracking, resource metrics visualization (CPU, memory, network), centralized log aggregation with search and filtering, and deployment controls for managing applications through a web dashboard. The platform features automated health monitoring with liveness and readiness probes, service discovery, intelligent load balancing, and horizontal pod autoscaling. Using the Kubernetes API client, Socket.io for real-time updates, and Material-UI for the interface, it demonstrates cloud-native development practices and modern DevOps workflows with integrated Prometheus and Grafana monitoring.

Automated DevOps platform implementing GitOps workflows for portfolio applications. Features CI/CD pipelines with GitHub Actions for automated build, test, and deployment processes, automated Kubernetes deployments with rollback capabilities, and infrastructure-as-code management with Helm charts. The platform provides centralized management of all portfolio applications with Docker registry integration, post-deployment health monitoring, and real-time status tracking. Built with Node.js, React/Vite, and Express, it includes a GitOps sync service that continuously monitors the repository and automatically applies changes to the Kubernetes cluster. With support for multiple environments (dev, staging, production), automated rollback on failure, and integration with Prometheus and Grafana for monitoring, PodRick demonstrates modern DevOps best practices including immutable infrastructure, semantic versioning, and blue-green deployments.

Multi-tenant Kubernetes platform enabling isolated deployment of portfolio applications in separate namespaces. Ideal for client demos, A/B testing, and staging environments. Features namespace isolation, resource quotas, and per-tenant monitoring. Built with React, Vite, Node.js, and Kubernetes, the platform provides isolated instances of educationELLy, educationELLy GraphQL, Code Talk, Bookmarked, FireBook, and IntervalAI applications running in different namespaces. Each tenant environment is completely isolated with its own resources, ensuring secure and efficient multi-tenancy. Utilizes Helm charts for deployment management, Nginx Ingress for routing, and Docker for containerization, demonstrating cloud-native architecture patterns and Kubernetes best practices for multi-tenant applications.

Enterprise-grade microservices platform demonstrating platform engineering and service-oriented architecture. Currently implements authentication and analytics services with plans to orchestrate educationELLy, Code Talk, FireBook, and IntervalAI applications. The auth service provides JWT-based authentication with PostgreSQL and Redis for session management, while the analytics service captures events with Kafka streaming and stores time-series data in InfluxDB. Built with Node.js and Express, the platform uses Docker Compose for local development with MongoDB, Redis, Kafka, Zookeeper, and InfluxDB infrastructure services. Features shared platform modules for logging, error handling, and middleware, demonstrating microservices patterns including service decomposition, event-driven architecture, and centralized authentication. The project showcases domain-driven design, API development best practices, and modern DevOps workflows with Docker containerization.