Inspiration

Attending networking events is high-value but overwhelming — hundreds of attendees, limited time. meetrank automates the entire research and prioritization pipeline so you can walk in knowing exactly who to find.

What it does

meetrank takes a Luma event you've RSVPed to and your networking goals, then returns a ranked shortlist of the people most worth meeting. It scrapes attendee profiles from Luma, pulls their LinkedIn data, summarizes each profile, and uses a pairwise LLM ranking system to surface your top contacts — each with a one-sentence explanation of why they ranked highly.

How we built it

  • Python backend paired with a mobile-friendly web frontend
  • Luma API to fetch attendee lists including linked social profiles
  • Bright Data LinkedIn scraper to pull structured profile data in parallel
  • LLM summarization to condense verbose profiles into concise, comparable markdown
  • Partial merge sort with pairwise LLM comparisons — O(A log N) comparisons instead of O(A²), using tool-calling to force deterministic picks with no tie ambiguity

Built With

Share this project:

Updates