While building Folio News, I faced several technical and product challenges. Scraping news from multiple websites was difficult due to different HTML structures, dynamic content, rate limits, and frequent layout changes, while still ensuring only authentic and relevant articles were collected. Designing a system to suggest news based on user likes was another challenge, as it required balancing personalization without creating echo chambers, efficiently tracking user preferences, and ranking content in real time. Additional challenges included maintaining data freshness, avoiding duplicate articles from different sources, generating accurate AI summaries without losing context, and scaling the system to handle increasing users and sources while keeping performance fast and reliable.