Speaker Curation Engine Deep-Dive

Our speaker curation engine balances three inputs: topic relevance, diversity (gender, ethnicity, role), and credibility (real-sounding titles and companies).

IEEE's conference guidelines ↗ and ACM's diversity statements ↗ inform our approach — conferences that reflect their audiences perform better. For AI generation, we prompt for "industry leaders, academics, and practitioners" with "gender/ethnic diversity." For fallback mode, we maintain a database of 60 speakers across 6 topics — AI, Web3, Climate, Health, Fintech, General. Each has a name, role, and professional photo from our Unsplash pool.

We never reuse the same speaker twice in one event. Fisher-Yates shuffle picks 8, then we assign photos. The result feels curated, not random. Fourwaves ↗ uses similar heuristics for academic conferences — the principle is the same: match speakers to the event's audience and goals.