Inusha Gunasekara, an IT undergraduate at the University of Moratuwa, built a cyberpunk-style terminal portfolio for Google's AI challenge. Dubbed "System 1," it uses Next.js, Tailwind CSS, and Google Cloud Run to create a digital twin. Instead of a standard landing page, it presents his skills through a command-line interface.

To ensure reliability, he pivoted from a live Gemini 2.5 Flash-Lite API connection to a deterministic simulation engine. This local mock stream delivers the sentient AI feel with near-zero latency and no ongoing costs. The architecture prioritizes a production-grade user experience, avoiding potential API quota issues or downtime for public access.

Key features include a hidden "kill switch" command for system crashes and a secure server-side proxy for live Ethereum gas tracking. The project highlights a common engineering trade-off: removing dependencies for stability. By building a local simulation, Gunasekara created an engaging, performant portfolio that showcases both frontend polish and backend pragmatism.