PhD student, MIT CSAIL · Recursive Language ModelsPhD student, MIT CSAIL · Recursive Language Models
nlpsystems
MIT CSAIL PhD student advised by Omar Khattab and Tim Kraska, working between NLP and ML systems. First author of Recursive Language Models (December 2025): instead of stuffing a long prompt into the context window, the model treats it as an external environment it can inspect programmatically, decompose and recursively call itself over — handling inputs up to two orders of magnitude beyond the context window and beating common long-context scaffolds even on shorter ones. The paper was accepted to NeurIPS 2026. Also works on GPU kernels and multi-agent systems, around a common thesis: better harness design unlocks capability already latent in existing models.
MIT CSAIL PhD student advised by Omar Khattab and Tim Kraska, working between NLP and ML systems. First author of Recursive Language Models (December 2025): instead of stuffing a long prompt into the context window, the model treats it as an external environment it can inspect programmatically, decompose and recursively call itself over — handling inputs up to two orders of magnitude beyond the context window and beating common long-context scaffolds even on shorter ones. The paper was accepted to NeurIPS 2026. Also works on GPU kernels and multi-agent systems, around a common thesis: better harness design unlocks capability already latent in existing models.