GLEMAD · AI SECURITY RESEARCH INSTITUTION
Glemad is an AI security research institution developing the Ollandi model class to understand live infrastructure, interpret attacker intent, support controlled response, and preserve evidence as events unfold.
THE RESEARCH QUESTION
Modern attacks move between infrastructure domains faster than fragmented tools can explain them. More automation alone does not solve that problem; action without context can become another source of risk.
Our work joins continuous state, threat interpretation, policy, bounded response, and verification into one inspectable discipline.
LATEST MODEL · PREVIEW
Our most capable security reasoning model yet, and the intelligence powering the Ollandi product.
Ollandi 5 Preview maintains a live model of cross-domain infrastructure, tests competing explanations for change, evaluates response against authority and consequence, and preserves the evidence behind every decision.
Read the technical paperRESEARCH PROGRAMS
Research is organized around the decisions a defensive system must make and the evidence required to trust those decisions.
We develop security-native models that maintain state across cloud, identity, runtime, network, endpoint, and private infrastructure while evidence and adversary behavior continue to change.
Enter the programOur evaluations test reasoning integrity, adversarial resilience, uncertainty, policy adherence, reversibility, and the point at which a system must return control to people.
Enter the programGlemad studies adversary activity, infrastructure exposure, regional conditions, and the consequences that matter to institutions carrying systemic risk.
Enter the programWe investigate how defensive systems can propose and verify the smallest effective response without exceeding explicit policy, operator authority, or service constraints.
Enter the programSAFE AUTONOMY
Glemad treats boundaries, evaluation, and evidence as part of model capability, not controls added after deployment.
Review our safety work ↗SELECTED PUBLICATIONS
The architecture, reasoning system, and control model behind Glemad’s latest security reasoning model.
A regional assessment of adversary activity and infrastructure exposure across consequential African institutions.
Practical guidance for strengthening regulated infrastructure while interpreting localization obligations.
RESEARCH IN OPERATION
Ollandi turns Glemad’s model research into a controlled cyber-defense product for teams responsible for real infrastructure. It connects live state, threat reasoning, policy-checked response, and verified evidence.
Explore Ollandi