Sujal Raj

Built a scalable template engine in React and Node.js enabling real-time content customization, cutting admin overhead by 40%. Shipped production-grade UIs in Tailwind that lifted user engagement by 30%, and hardened the backend by refactoring legacy REST endpoints with robust error-handling middleware — tightening CI/CD cycles by 20%.

An AI-powered workspace that turns any document collection into a queryable knowledge system. Built an autonomous GraphRAG ingestion pipeline with LangGraph that fuses semantic vector search (Qdrant) with a relational knowledge graph (Neo4j) to enable complex cross-document reasoning. The Next.js interface renders interactive ReactFlow visualizations of extracted entities, while Mem0 provides long-term conversational memory for personalized answers. Backend is a production-ready FastAPI service deployed to AWS EC2 via Docker + GitHub Actions, with S3 for scalable document storage and async background processing.

Full-stack analytics pipeline that scrapes and analyzes 600+ product reviews across 6 brands, with ASIN-level caching and automated retries for 100% data recovery. Behind it sits a hybrid AI insight engine using GPT-4o-mini with async batching to extract aspect-level sentiment across 10 dimensions — surfacing data-backed Recommended Actions and automated brand rankings via a FastAPI REST layer.

A real-time collaborative IDE with multi-user code editing, live file-tree sync, and in-browser execution via the WebContainer API — achieving 100% state sync across concurrent sessions. Scaled with a Socket.IO + Redis architecture and integrated Gemini for contextual code suggestions and @ai-mention workflows, secured with JWT auth.