RiderEx
When a passenger submits feedback, 5 AI agents collaborate through Band to classify the issue, triage safety, draft a response, quality review it, create an engineering ticket, and render a Go/No-Go software release decision — all in one pipeline run. What RiderEx Does: A passenger reports a problem with their autonomous vehicle ride. RiderEx: -Classifies the feedback (safety, comfort, route, software, or compliment) -Triages safety risk and NHTSA reportability -Drafts a warm, policy-compliant customer response with refund/credit -Quality reviews the response for tone, legal exposure, and recurring patterns -Creates a JIRA engineering ticket and assigns it to the right team -Renders a Go / Conditional / No-Go release decision for the vehicle's software branch -Saves the full record to Supabase and updates every tab in real time
Category tags:
"Application of technology: Excellent implementation of 5 Band agents in a sequential pipeline (intake, safety triage, resolution, quality review, engineering handoff). Clean use of Claude via AI/ML API with proper Band SDK integration and Supabase persistence. Presentation: Clear demo video and well-structured README with architecture diagrams, tab walkthroughs, and API endpoint documentation. The live Vercel demo adds strong credibility. Business value: Real-world AV customer support automation with NHTSA compliance, JIRA integration, and Go/No-Go release decisions. Strong enterprise applicability for autonomous vehicle operators. Originality: Innovative combination of multi-agent pipeline with AV-specific safety triage and regulatory reporting. The 250k seeded dataset and fleet management UI show depth beyond a typical hackathon submission."
Sanem Avcil
For Innovators & Creators