Security engineer at ControlPlane. I work on agentic security in regulated industry, mostly fintech: building the platform agents run on and red teaming them in a sandbox before they reach production.
Open the nav computer to explore the chart. The same work is listed below.
- GPT-4o (Pre-release model, 2024): System card, Case study
- Operator (Computer-using agent, 2025): System card, Case study
- o3-mini (Reasoning model, 2025): System card, Case study
- Deep Research (Research agent, 2025): System card, Case study
- Wild West I (Blog post): Read it
- Wild West II (Blog post): Read it
- Agent runtime (Platform engineering): Build and operate the runtime AI agents execute on at financial services clients: sandboxed execution, scoped identities and tool permissions, and the SRE work that keeps agent workloads observable and recoverable on Kubernetes.
- AI Security training (Course author and lead trainer, 2026): Author and lead ControlPlane's AI Security training. The agent sandboxing content was built with Docker, who sponsored the 2026 North America roadshow, starting in Toronto and New York.
- A Game of Ethics (Open source eval harness, 2026): Repository
- DynastAI (GRPO demo): Repository
- FINOS AI Governance (Industry framework): Framework
- FINOS AI Readiness (Industry framework): Repository
- AI Governance Framework (Open source): Repository
- VEX remediation (Supply chain security): Scanners produce more findings than any team can triage. I operationalize VEX so a finding comes with a statement about whether it's exploitable in that build, then use agentic remediation and sandboxed red teaming to work the ones that are.
- Argo CD threat model (CNCF end user threat model, 2023): Threat model, Case study
- Repo-server RCE (Write-up, 2026): Read it
- Envoy Gateway threat model (CNCF end user threat model): Threat model, Blog post
- EKS at Citi (Security architecture): Case study
- SPIFFE and SPIRE (Linux Foundation course, LFS482): Course, Case study
- Intro to Zero Trust (Linux Foundation course, LFS183): Case study
- High assurance at NIST (Conference write-up, 2023): Read it




