Brooks McMillin
AI Security Researcher & Machine Learning Protection Specialist
I develop cutting-edge security solutions for AI/ML systems, focusing on adversarial machine learning defense, LLM security, and protecting production AI infrastructure. Currently at Dropbox, I lead initiatives in securing large-scale language models and building robust defenses against emerging AI threats.
Defense in Depth for AI-Assisted Development: Pre-commit Hooks, Review Agents, and CI That Catch LLM Mistakes
Practical strategies for safer AI-assisted development: automated review agents, layered security checks, and context management that prevents catastrophic mistakes.
Read ArticleCurrent Focus
Advanced AI Security Research
Leading research in protecting AI systems from adversarial attacks, prompt injection, and model extraction. Developing next-generation security frameworks for LLMs, multimodal AI, and autonomous AI agents.
- Building real-time detection systems for adversarial inputs and prompt injection attacks
- Developing security benchmarks for evaluating LLM robustness and safety
- Researching privacy-preserving techniques for federated learning and distributed AI
- Creating defensive strategies against model inversion and membership inference attacks
Featured Projects
Agent Framework
productionOverview
A production-ready framework for building LLM agents with the Model Context Protocol (MCP). This framework provides the foundation for …
- Production-ready framework for building LLM agents with MCP
- Full OAuth 2.0 with PKCE and dynamic client registration
Secure Task Management Suite
productionOverview
A comprehensive secure task management platform consisting of three integrated components: a main application, Python SDK, and MCP server …
- Full OAuth 2.0 authorization server with PKCE support
- Comprehensive security testing suite with Vitest
Secure Communications Suite
productionOverview
A dual-language secure communications suite providing SMS capabilities through GSM modems, built with both Go and Python implementations. …
- Cross-platform GSM modem interface (Go + Python)
- Real-time SMS communication with security considerations
Recent Appearances
Panel discussion on how AI-driven threats evolved in 2025 and what defenders should prepare for in 2026
Technical deep-dive on implementing LLM security controls at scale using Lakera Guard for Dropbox's AI features
Let's Connect
Interested in AI security research, collaboration opportunities, or speaking engagements? I'd love to hear from you.
More About Me