About
Aleph Beth is an independent practice dedicated to AI-agent security. We help companies and institutions understand what their agents can do, test it the way an adversary would, and put in place controls that hold.
Expertise
Mathematics
Understanding models from the inside: what a transformer can do, what it never guarantees, and why.
Cybersecurity
Modelling threats, testing like an attacker, designing deterministic controls outside the model.
Artificial intelligence
Agent architectures, tools, memory, document retrieval: where the risk moves when AI takes action.
Approach
- We only test within a scope authorised in writing.
- We do not sell compliance: we deliver findings, evidence and recommendations.
- We publish our research, because a fast-moving field defends itself better in the open.
- We say what we do not know yet, including about our own product.
Languages
Engagements in French and English, in France and remotely worldwide.
Status
Company in formation — engagements open in January 2027.
Request a security review
Describe the agent, its access and the business decision it can influence. We reply within two business days.
Request a review