Verification, composition and evolution for LLMs and agents: typed operad workflows, checked rewrites, exact mathematical certificates, an epistemic immune system, prompt lineage, adversarial swarms, budgeted execution and replay. Python/Rust runtimes, Lean 4 kernels, C# tools, research and teaching.
python rust ai-safety operads formal-verification-methods prompt-engineering llms llm-evaluation-framework agent-skills multicategory llm-manipulation ai-safety-research llm-sparse-reasoning
-
Updated
Oct 31, 2026 - Rust