Zeitgaist

Decision-makers in finance, marketing, and research need to understand public opinion across Twitter/X, Reddit, Hacker News, Mastodon, Bluesky, and other platforms. Manual monitoring is time-consuming, critical insights in non-English sources are missed, and traditional search lacks temporal context and source attribution. Ask "why did real estate prices surge recently?" against Chinese sources and you get completely different answers than English - geopolitical drivers, local policies, cultural factors that Western media doesn't cover.

Built a unified backend serving two complementary products - an AI chatbot for conversational queries and an analytics dashboard for trend visualization. The production Hybrid RAG system features two-stage retrieval (dense embedding search + cross-encoder re-ranking) — a Corrective RAG pattern that improved relevance@10 from 72% to 89% compared to dense-only retrieval. Initially tried single-stage retrieval but found accuracy degraded on cross-lingual queries - the two-stage approach with language-specific re-ranking solved this at acceptable latency (~200ms p95).