How we made AISQL up to 8x faster and 70x cheaper through AI-aware planning, adaptive model cascading, and semantic join rewriting.
Benjamin Han
Benjamin Han is a Staff Software Engineer at Snowflake AI/ML, where he has designed and launched several LLM-powered features in the AISQL suite, including CLASSIFY_TEXT, AI_FILTER, AI_AGG, and AI_SUMMARIZE_AGG. He is currently focused on improving batch query performance through Online Fine-Tuning. Prior to Snowflake, Ben led several projects in Reinforcement Learning and Multi-Armed Bandits at Lyft, with research published at ACM SIGKDD and INFORMS IJAA, and was named a 2023 INFORMS Franz Edelman Laureate. Based in San Francisco, he spends much of his free time outdoors and training towards becoming a certified Mountain Guide.
NOV 04, 2025|8 min read
