Software Engineer & Researcher

Secure AI Systems for Agentic LLMs.

I build infrastructure for tool-using LLM agents: efficient memory/serving, agent orchestration, and least-privilege security. Software Engineer at Ona (formerly Gitpod). OpenFGA maintainer.

Software Engineer @ Ona (formerly Gitpod)
OpenFGA maintainer (Google Zanzibar-style authorization)
Building OSS: TokenVM, Agentflow, LLMTraceFX, A2AS, actionsec

Research Statement Resume Contact

Featured Work

Research prototypes and open-source contributions

View all projects

TokenVM

High-performance runtime treating LLM KV cache as virtual memory across GPU VRAM, RAM, and NVMe

ai-systemsdistributed-systems

Intelligent paging
Compute-copy overlap

Agentflow

Kubernetes for AI agents - orchestration, security, observability, and cost-aware scheduling

agentsdistributed-systemssecurity

DAG execution
Multi-provider routing

OpenFGA

Google Zanzibar-style fine-grained authorization system (maintainer)

authorizationdistributed-systemssecurity

Performance optimization
Correctness fixes

View all projects

Research Focus

Secure KV-cache memory, agent authorization, and adversarial robustness

Learn more

Open Source

OpenFGA maintainer, GitHub1s contributor, and OSS builder

Explore

Writing

Technical articles on Dev.to, Medium, and research essays

Read