An Intelligent AI-Driven Platform for Automated Portfolio Generation and Real-Time Publishing

Abstract

This paper presents Portiva, a cloud-based Software-as-a-Service (SaaS) platform designed to automate the generation, deployment, and management of professional portfolio websites. The system collects structured user data through a guided multi-step onboarding process and utilizes a locally hosted Large Language Model (LLM) via Ollama to generate high-quality, context-aware content. The generated portfolio is dynamically rendered and published on a unique public URL within minutes.

The primary contribution of this work is a schema-driven placeholder mapping architecture, where each template explicitly defines its data requirements using a structured placeholder_schema stored in the database. The AI engine interprets this schema to generate only the required content fields, eliminating hardcoded mappings and enabling seamless extensibility of templates without modifying application logic.

The platform is implemented using a modern full-stack architecture comprising React 19, Spring Boot 3.2, FastAPI, PostgreSQL, and Redis. It incorporates production-grade features such as asynchronous AI processing, Redis-based caching, soft deletion, and subscription management via Razorpay. Experimental evaluation demonstrates sub-200ms API response times and complete end-to-end portfolio generation within three minutes on commodity hardware, highlighting the system’s efficiency and scalability.