MeetMatch

Inspiration

Today’s event made something clear: most people walk into meetings—online or in-person—without knowing who they should talk to. Networking becomes accidental, and introverts or newcomers often struggle the most. I wanted a system that removes uncertainty and makes meaningful connections intentional, personalized, and effortless.

What it does

MeetMatch AI analyzes participants, builds Networking Personas, and identifies the top people each attendee should meet. It explains why the connection matters, offers personalized icebreakers, and updates matches in real time as new participants join. After the event, it learns from user feedback to refine future matchmaking.

How we built it

We designed an agent-based workflow using Parallel, Sanity and ClaudeAPI.

PersonaCreator Agent generates structured participant profiles.
MatchGenerator Agent computes match scores.
SelfImprover Agent adjusts weights using feedback.

Redis stores personas, matches, and feedback. We integrated a LinkedIn-compatible API to reliably access participant information.

Challenges we ran into

The hardest challenge was obtaining clean participant data. Direct LinkedIn scraping is unreliable and restricted, but an API finally solved the data-access bottleneck. Another challenge was designing icebreakers that feel natural and genuinely helpful for introverted users.

Accomplishments that we're proud of

We built a fully autonomous, self-improving networking agent that takes action—not just generates insights. The system learns from every interaction and produces higher-quality matches with each event.

What we learned

Building matchmaking logic is closer to an evolving optimization problem than a static recommendation system. We learned how to encode user intent, design iterative scoring rules, and build a reinforcement loop using real-world feedback.

What's next for MeetMatch

We plan to support Slack/Discord-based events, add group-match recommendations, and build a predictive model that forecasts high-value connections. Long-term, MeetMatch could become a universal layer that optimizes every meeting, conference, or online interaction.