Threshold k-of-n oracle aggregating TEE attestation quotes (TDX/SEV-SNP/Nitro) across heterogeneous vendors for AI-liability parametric insurers.
Store how an agent's behavior changed in response to user corrections as a first-class, versioned, replayable artifact — not facts.
Memory-first Rust/Go agent runtime with built-in signed append-only audit logs that generate EU AI Act Annex IV / SR 11-7 compliance artifacts end-to-end.
Tooling for the hard 60-80% of eval-dev time (Hamel Husain) — authoring and maintaining failure taxonomies — which no tool owns.
Apply GEPA (genetic Pareto reflective optimizer) to LLM router policies — text artifacts with scalar metrics — not just to prompts.
A governance/security layer for MCP servers — 43% have OAuth flaws, 43% command-injection, 22% out-of-scope access, and no dominant vendor exists.
Poisoning red-team-as-a-service for frontier model training corpora — the training-time attack surface no one else covers.
Cryptographically signed corpus manifests + statistical trigger-phrase detectors + EU AI Act Art. 53(1)(d) artifact authoring — C2PA-for-text.