Description
Summary Description: Grit includes a team messaging feature ("Circles") with group and direct conversations, plus an automated content-moderation pipeline that's unusually well-specified: every message runs through a moderation API, gets classified into one of 15 named flagging reasons across three severity tiers, and routes to the right reviewer automatically based on both the flag severity and the sender's role. We need this built for real, replacing the mock data in our current prototype. Scope: Real-time messaging (group + 1:1), read receipts, message history Integration with a content moderation API (OpenAI Moderation endpoint or Perspective API — your call on which fits better) The tiered flagging/escalation logic exactly as specified (this is the most detail-dense part of the spec — read §5.3 and §5.3.1 of the architecture doc closely before bidding) Structured "Assignments" feature living inside message threads (coach-assigned tasks, athlete submission, completion tracking) Required experience: Strong experience with real-time messaging infrastructure — WebSockets, or a managed provider like Stream Chat/PubNub if you can make the case for it Prior work integrating a content moderation/trust-and-safety API into a product Comfortable implementing conditional business logic with several interacting rules (role-based routing, severity tiers, auto-escalation thresholds) — this isn't simple CRUD Bonus: any prior experience building products with real child-safety considerations (this platform has users under 13) Deliverables: Working messaging system integrated with the mobile app team's build, moderation queue data flowing correctly to both team-level and platform-level review surfaces.