developer

FrontendForge is designed as a zero-setup coding environment for developers, students, and educators. Each active session receives an isolated sandbox with a Vite development server and a terminal-capable agent, while the browser provides a VS Code-style workflow.

The resume describes the project as a five-microservice cloud IDE covering authentication, sandbox provisioning, AI orchestration, routing, and notifications.

• On-demand, isolated per-user React + Vite sandboxes inside Kubernetes.
• VS Code-style browser interface with file explorer, live preview, terminal, and AI chat.
• LangGraph ReAct agent powered by Mistral Large that can list, read, create, and update project files.
• Server-Sent Events stream AI activity and final responses back to the browser.
• Vite HMR reflects AI-written file changes in the live preview without a full page reload.
• NGINX Ingress and a custom WebSocket proxy route preview and terminal traffic by sandbox ID.
• RabbitMQ event pipeline handles login notification delivery, while Google OAuth 2.0 and JWT protect authentication.

Kubernetes Client API provisions pods and services. Redis tracks active sandbox state and expiry. Shared volumes let the agent write files that Vite immediately observes. Socket.IO provides terminal communication, while SSE keeps AI generation streamed to the UI.

The resume also highlights an intended sandbox startup target under five seconds and AWS S3-backed project-file persistence as part of the project architecture.

Snitch covers the complete commerce journey: authentication, catalog browsing, product detail, search, filtering, sorting, cart, wishlist, checkout, payments, orders, and an administrative panel.

The resume describes it as an end-to-end production-ready e-commerce platform and highlights performance-focused product search, filtering, sorting, pagination, and Redux Toolkit state management.

• JWT authentication, Google OAuth 2.0, email/password login, HTTP-only cookies, protected routes, and RBAC.
• Product browsing with details, search, filtering, sorting, cart quantity updates, and wishlist support.
• Razorpay payment integration with payment verification and secure order placement.
• User order history, order details, dashboard, and administrative order-status management.
• Admin dashboard for products, categories, orders, and users.
• MongoDB Atlas with Mongoose, aggregation pipelines, and optimized queries.
• Responsive interface with loading states, toast notifications, and Cloudinary image delivery.

The project uses MongoDB aggregation pipelines for data-heavy API work and Redux Toolkit for predictable client-side state. The resume specifically highlights server-side Razorpay signature verification to protect payment flows and optimized pagination for product discovery.

The repository README documents Vercel for the frontend, Render for the backend, and MongoDB Atlas for the database.

Blinkly AI combines conversational AI, real-time web search, and a multi-model Battle Arena in one application. Users can maintain persistent conversations, retrieve current information, and compare model responses through an AI-powered evaluation workflow.

The project README describes the central problem as single-model AI applications not always producing the best answer. The arena addresses that by comparing multiple model responses and making the evaluation process visible.

• AI-powered conversations with real-time responses, persistent chat history, and session restoration.
• JWT authentication, protected routes, session management, and user-specific chat history.
• AI Battle Arena comparing Mistral and Groq responses with Gemini acting as the judge.
• LangGraph multi-agent workflow with AI-generated scoring and automatic winner selection.
• Tavily-powered real-time web search for up-to-date information retrieval.
• Responsive desktop and mobile experience.

The documented architecture is React frontend → Express backend → LangGraph workflow → AI models → response evaluation → final result. The frontend is organized around auth, chat, battle, and pricing features, while the backend separates controllers, routes, models, middleware, services, sockets, and config.

• Full-stack application architecture and production deployment.
• AI application architecture, LangGraph workflows, and multi-agent systems.
• Authentication and authorization with JWT and protected routes.
• Redux state management, API integration, session restoration, and real-world debugging.