Kevin Zhang

Third-year at UC San Diego researching multi-agent alignment. Thinking about the theology of AI risk.

I'm a third-year undergraduate at UC San Diego double majoring in Data Science and Math-CS. I work in Babak Salimi's machine learning lab, where my current research applies recommender systems to the AI tool ecosystem at scale.

I care deeply about reducing catastrophic risk from advanced AI. Beyond research, I co-organize UCSD's AI Alignment Club and am building connections across the safety ecosystem through Blue Dot Impact, Action Potential, and Apart Research.

Outside the lab, I'm a Reformed Christian involved in campus ministry, an avid D&D game master, and a daily swimmer. My long-term vision is founding an ecumenical nonprofit bridging AI safety and Christian stewardship.

Current Work
ToolFlix: Why the MCP Ecosystem Needs a Recommender System
Agents can't find the right tools. We built a recommender system that learns from execution feedback to surface tools that actually work, not just tools that match a description.
Read more →
Simulating Moltbook: How Fake Peers Extract Secrets from AI Agents
Testing how much a network of fake peers can extract from an agent, built on the InspectMAS benchmark.
Read more →
Past Projects
ToolTrace: Fix the Environment, Not the Agent
When agents fail at tool use, fix the tools, not the model. An automated loop that analyzes failed traces and generates code-level wrappers to prevent recurring failures.
Read more →
Apart Research Hackathon: Open-Weight Model Safety
AI safety hackathon submission exploring evaluation and alignment challenges for open-weight models.
Read more →
Latent Adversarial Training Replication
Replication and extension of latent adversarial training techniques for improving LLM robustness.
Read more →
Evaluating Genomic and Clinical Risk Factors for Alzheimer's Disease in Individuals with Hypertension
Elizabeth Kim, Kevin Zhang, Miski Abdi, Wei Tse Li, Ruomin Xin, Jessica Wang-Rodriguez, Weg M. Ongkeko
Biomedicines, 2025
Read more →
Rebuilding the Tower
What Babel teaches us about AI alignment.
Read more →
Going to Action Potential
A retelling of my experience at Action Potential, an AI safety retreat for university students organized by Kairos.
Read more →