Uploaded July 2026 | Updated September 2026, 2 weeks ago
SEAR is a schema-based system for evaluating production LLM responses and routing requests across providers in LLM gateways, defining an extensible relational schema spanning quality signals and operational metrics across roughly one hundred SQL-queryable columns. Because signals come from LLM reasoning rather than shallow classifiers, SEAR supports interpretable routing and large cost reductions at comparable quality across thousands of production sessions.
Speakers & affiliations: Zecheng Zhang (Strukto.AI), Han Zheng (Infron.AI), Yue Xu (Infron.AI)
Session: Paper Session 8 — AI Systems in Practice · Friday, May 29
Key terms: LLM gateways, evaluation, routing, schema, cost reduction, production systems
ACM Digital Library: doi.org/10.1145/3786335.3813131
SEAR is a schema-based system for evaluating production LLM responses and routing requests across providers in LLM gateways, defining an extensible relational schema spanning quality signals and operational metrics across roughly one hundred SQL-queryable columns. Because signals come from LLM reasoning rather than shallow classifiers, SEAR supports interpretable routing and large cost reductions at comparable quality across thousands of production sessions.
Speakers & affiliations: Zecheng Zhang (Strukto.AI), Han Zheng (Infron.AI), Yue Xu (Infron.AI)
Session: Paper Session 8 — AI Systems in Practice · Friday, May 29
Key terms: LLM gateways, evaluation, routing, schema, cost reduction, production systems
ACM Digital Library: doi.org/10.1145/3786335.3813131
