Mohammed Maqsood

Built and deployed demo shortshub.app to experiment coding with LLMs — a full-stack short-form video ideation and growth intelligence platform — using a manifest-driven kernel architecture that enables AI coding agents to develop features in safe isolation. Core capabilities include AI trend analysis (Viral Radar), automated video editing, thumbnail extraction, platform safety compliance checking. Includes an Admin Control Plane for runtime feature flag management, user tier administration, and auto-scanned implementation documentation across the codebase. Stack: Typescript, PostgreSQL, Tailwind CSS on Google Cloud Platform with GCS for video storage. AI powered by Google Gemini (Vertex AI) with Groq as a zero-cost failover. Built entirely using AI coding agents.