AI Researcher & Systems Engineer
Multi-Agent Systems • LLM Reliability • Verification & Causal Failure Attribution

I am an undergraduate researcher in Artificial Intelligence working on making autonomous, multi-agent LLM architectures reliable, diagnosable, and verifiable.

• 🧩 Multi-Agent Systems & Verification: Investigating error propagation across inter-agent communications and isolating root-cause failures in collaborative agent networks.
• 📐 Evaluation & Benchmarks: Developing diagnostic evaluation protocols that identify where and why an agentic pipeline breaks down, moving beyond superficial end-to-end accuracy metrics.
• 🧠 Collaborations: Research collaborator on Theory of Mind for Multi-Agent Systems (ToMAS) with Toronto Metropolitan University.
• 🎓 Teaching & Mentorship: Section Leader for Stanford University's Code in Place, mentoring students through core algorithmic and computational thinking.

• 📄 llm-fallback-reliability: Reproducible benchmark and evaluation framework for fallback strategies in multi-agent LLM systems (Accepted to NeurIPS Workshop).
• 🔍 counterfactual-communication-attribution: Causal communication attribution in multi-agent networks to determine which message broke the system.

Open to international research internships, visiting researcher opportunities, and collaborations in reliable agentic AI.