A software engineer submitted a portfolio for Google AI's 'New Year, New You' challenge, showcasing two production-ready systems built on Google Cloud. The projects address common developer pain points: information overload from RSS feeds and monitoring AI code review pipelines. Both solutions leverage Gemini 2.0 Flash for intelligent processing, demonstrating practical AI integration in daily workflows.

The Content Intelligence Hub tackles article curation by combining Gemini relevance scoring with community validation from HackerNews and Lobsters. This dual-scoring algorithm filters 500+ daily RSS items into ~10 high-value reads, solving the problem where AI alone can't distinguish obscure posts from battle-tested engineering content. The system runs on Cloud Run and BigQuery for under $20 monthly.

The LLM Code Review Observability project provides real-time monitoring for AI-powered PR reviews. It tracks RAG retrieval quality, cost trends, and error rates via live dashboards querying BigQuery directly. A five-layer security model with BudgetGuard caps spending at $2/day, using Firestore for rate limiting. This offers actionable insights for tuning review systems across different codebases.