Shubham Jain

The paper introduces CONCORD, a privacy-aware framework that enables proactive, always-listening AI assistants to safely recover necessary dialogue context through negotiated exchanges, thereby addressing core privacy risks.

The paper proposes a novel approach for solving Bit Manipulation Puzzles in the NVIDIA Nemotron Model Reasoning Challenge by abandoning arithmetic logic in favor of string similarity, structured search, and autonomous error recovery.

This paper presents Roomie, an architecture for orchestrating model serving in multi-model scenarios on GPUs, reducing SLO violations by predicting and avoiding kernel-level interference.