Lineage: Evidence-Linked Due-Diligence Drafting with Bounded Model Authority
Source-linked due-diligence drafting, citation checks, evaluation results and human approval boundaries.
Research
Evaluation and evidence in AI systems
Technical reports on evaluating generated infrastructure configurations and building evidence-linked research workflows.
IaC Eval Bench: Auditing Static Evaluators for LLM-Generated Kubernetes Configurations
A static Kubernetes evaluation harness, a scoring audit and a reproducible comparison of saved model outputs.