Shayan Talaei

Position: The Hidden Costs and Measurement Gaps of Reinforcement Learning with Verifiable Rewards

SPRINT: Enabling Interleaved Planning and Parallelized Execution in Reasoning Models

Reasoning-SQL: Reinforcement Learning with SQL Tailored Partial Rewards for Reasoning-Enhanced Text-to-SQL

CHASE-SQL: Multi-Path Reasoning and Preference Optimized Candidate Selection in Text-to-SQL