Tutorials

Hands-on, reproducible guides. Everything I publish here is something I actually run myself.

Technical material available now

• Building a 23-GPU local LLM inference cluster

  Architecture notes on ROCm, multi-node operation, repeatable Python benchmarks, and SQLite result storage.

• Evaluating prompt injection in tool-using LLM systems

  A practical workflow for threat models, attack corpora, traces, scoring, and regression tests.

In the pipeline

• Running local LLMs on AMD GPUs with ROCm

  From bare metal to vLLM serving, including the pitfalls.

• Automating security workflows with n8n

  Repeatable evaluation runs and compliance checks as pipelines.

• Building a prompt injection test harness

  Generate, execute, and score adversarial test cases against your own LLM stack.