news

Sep 2026 MonarchRT was accepted to NeurIPS 2026!
Jul 2026 Released FlashRT, an agent harness that guides coding agents to turn simple reference implementations of real-time multimodal applications into optimized multi-GPU deployments.
Feb 2026 Released MonarchRT, a highly effective sparse attention parameterization for real-time video generation.
Jan 2026 Teaching assistant for 18-789: Deep Generative Modeling at CMU this spring.
Aug 2025 Started my PhD at CMU in InfiniAI Lab, advised by Prof. Beidi Chen!
Jun 2025 L3M+P was accepted to IROS 2025!
May 2025 Started a research internship at NVIDIA, working on efficient MoE serving.
May 2025 Graduated from UT Austin after defending my undergraduate honors thesis!
May 2024 Interned at IBM Research in Yorktown Heights. Some of our work was published on the PyTorch blog!