Presentations

Over the years I’ve had the privilege of presenting some of my work at a variety of different conferences. I’ve collected information about some of them here.

Strange Things are Afoot: Using Faucet Coprocessing for Deception and Active Defense

Presented at CircleCityCon 8.0.

This is one of my favorite demos that I’ve ever done. It involved so many different moving pieces and I’m amazed that it held together as well as it did. This demo was really only possible because I was presenting remotely on my home network. The actual mechanics of it were really cool as well, there was a bunch of crazy L2 NAT happenening on the switch to make the swap outs seamless.

Defensive 5G

Presented with Eric Mair at DefCon 30 Demo Labs

Getting to present at DefCon was the thrill of a lifetime and getting to do so alongside Eric made it better. We presented a research project we had done wherein we used Software Defined Radio and containerized software to build out an experimental 5G base station that we used to do security testing. The whole thing fit into a pelican case a little bit bigger than a lunch box that I had to carry around the whole time. We showed phones communicating over it and everything.

Fun Fact: On the way back from DefCon on that trip, the TSA agent instructed me to put all electronics the size of a cellphone or larger into their own bin, I had 14 bins.

AI Assurance For System Operators

Presented at SecureWV 13

SecureWV is one of my favorite conferences to attend and present at. It is small but the community there is amazing! This aprticular talk was about my early experiences trying to do security testing on AI/ML enabled systems.

AI Enhanced Attacks: Model in the Middle

Presented at ShmooCon 2024 with Ari Chadda KernelCon 2024 with Ari Chadda SecureWV 15

This is, hands down, my favorite demo ever. I particularly love the VScode variant that we did for KernelCon where we intercept a prompt going from VSCode to an AI development assistant and have it instruct the user to use a library that I wrote in place of Pandas. The library just shows some ascii art of a bottle of Mellow Corn (KernelCon goes hard on the corn theme), but really it could have been anything.

Stochastic Garrottes: A Data Driven Approach to LLM Generated Malware

Presented at Wild West Hackin Fest 2025

I’m really proud of this bit of work. I think that it was a fundamentally sound experiment and the results were pretty fascinating. To a certain extent some of the things I was trying to do have been overtaken by some of the bigger labs, but I intend to revisit the construct here in the not too distant future using some newer tools and constructs to see if there has been any meaningful change in the space.