Cao Nguyen Pham

Research focus

Red-Teaming & Adversarial Robustness (2)Tool Use & Agents (2)Code Generation & Program Synthesis (1)Constitutional AI & AI Ethics (1)

Frequent co-authors

Oliver Bračevac (2)Yaoyu Zhao (1)Yichen Xu (1)Oliver Bravcevac (1)

Papers (2)

May 27, 2026

Yaoyu Zhao +8May 27, 2026·also BAIR

LACUNA: Safe Agents as Recursive Program Holes

What if you could let LLMs shape their own runtime, safely?

Yaoyu Zhao, Yichen Xu, Oliver Bračevac +6

Code Generation & Program Synthesis Red-Teaming & Adversarial Robustness Tool Use & Agents

Mar 1, 2026

Oliver Bračevac +1Mar 1, 2026

Tracking Capabilities for Safer Agents

Capability-based programming languages offer a surprisingly practical path to safer AI agents by statically preventing information leakage and malicious side effects during tool use.

Oliver Bračevac, Cao Nguyen Pham

Constitutional AI & AI Ethics Red-Teaming & Adversarial Robustness Tool Use & Agents

Search

Cao Nguyen Pham

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (2)