"AstroTree team on AI Infra Summit Hackathon Hackathon"
Team Idea
A congestion-aware LLM router. Instead of static quality and cost tables, it streams live TTFT, queue depth, and error telemetry from every backend, then routes each request to meet its latency SLO. Hedging and cascade fallback hold p99 through backend failures.
Team Members
.png&w=256&q=75)
Junghwan Kim
baobabtr33
Submission
Team Leader hasn't made a submission yet