Amin Hasbini

AI & Cybersecurity Executive

Amin Hasbini

Ex-Director, Kaspersky GReAT META Research Center. Expert contributor, French Senate OPECST report on AI (2024). Paris.

Available to build PQC and AI Agent Security capabilities inside institutions, or to contribute as external authority to existing programs. Role-creation conversations preferred.

NIS2, DORA, and the EU AI Act are pushing two new capabilities onto every European board agenda. I operationalize them with regulated enterprises, not as external compliance advice.

Pillar 01

Post-Quantum Cryptography Maturity

Crypto-agility for the post-quantum era.

Pillar 02

AI Agent Security

Authenticate, authorize, communicate for non-human identity.

Working with defense and aerospace groups, consulting firms, banks, and tech vendors.

  • Build. Authored the machine-learning threat-hunting research that delivered Kaspersky’s 25% APT detection lift in H1 2024 , across the 400 million users and 250,000 organizations protected by Kaspersky technologies. Built Kaspersky’s META threat intelligence practice. Attribution work on StoneDrill , MuddyWater , Desert Falcons , Gaza Cybergang , and Operation Ghoul .
  • Bridge. Expert contributor to the French Senate’s OPECST report on AI (2024) . Subject-matter expert on ICANN’s second DNS Security, Stability & Resiliency Review Team (SSR2, 2017-2019) , the global governance review of the internet’s root DNS infrastructure. Invited expert for Renaissance Numérique’s AI Dialogues (Geneva session alongside the Ministère de l’Europe et des Affaires étrangères, Brussels at UCLouvain). Speaker at 100+ international conferences. Featured on BBC, CNN, CNBC, Forbes, Bloomberg, France 24, Al Jazeera, Al Arabiya. Trilingual EN/FR/AR.
  • Field. 12 years at Kaspersky GReAT, 70 countries of regional threat coverage, engagement across 100+ countries, 34 with direct in-country travel. Security work across 10+ central banks (SAMA, Qatar CB, UAE CB, Lebanon CB), 25+ commercial banks, tier-1 oil & gas (ARAMCO as Kaspersky’s largest global customer, plus ADNOC, SABIC, Qatar Petroleum), major MENA telcos (STC, Etisalat, DU), and 6 international law-enforcement bodies (Interpol, Europol, Afripol, GCCPOL, FBI, ILCERT) including the FIFA World Cup Qatar 2022 joint operation with Interpol.

Current work

Five-paper non-human identity series complete. Latest: Auditing Agents Under NIS2, DORA, and the EU AI Act (May 19) and a joint LinkedIn essay with Cesar Cerrudo on AI offense, defense, and the governance vacuum (May 21). Series II on PQC Maturity opens late June 2026.

Writing → · Tools → · Track record → · Project document

Work together

Two ways I engage with institutions building PQC and AI Agent Security capabilities.

  • Builder inside. VP or Director of Innovation, named capability the board can point to, hand-off-ready program within 18 months. Best when the organization needs an internal owner who can ship, recruit, and answer to the executive committee.
  • External authority. Senior Advisor, Distinguished Fellow, or advisory-board contributor. Part-time, non-exclusive. Best when the program already has internal ownership and needs external intellectual authority, public voice, and a bridge to policy and threat-intelligence depth.

Role-creation conversations preferred. For current substance, see Papers and Track record .

Reference implementations

Nine working tools that demonstrate the PQC and AI Agent Security thesis in production:

  • Meetade : reference deployment of browser-native video conferencing on hybrid post-quantum TLS (X25519MLKEM768), cloud-portable and sovereign-migration-ready. Access by invitation.
  • PQC Scanner : reference scorecard measuring post-quantum readiness at CAC40 scale.
  • Agent-CBOM : crypto-agility inventory for AI agents, mapping each agent’s channel, identity, and receipt cryptography and scoring its post-quantum posture. Bilingual EN/FR.
  • Agent Identity Platform : reference implementation of the three-pillar governance flow (register, authorize, communicate) with auditable call graph.
  • AI Agent Security Maturity Assessment : reference diagnostic for agent identity, authorization, and communication maturity (three-pillar scoring).
  • AgentTrustLab : reference simulator for agent-to-agent authentication under PQC and ZKP constraints.
  • Predict-then-Permit : world-model trust layer that authorizes an agent on what it is about to do, not only on who it is.
  • DomainWatch : reference monitoring surface for domain-impersonation risk across 5,452 enterprises.
  • Papers Research Assistant (Beta) : live AAC reference: a scope-guarded, citation-grounded conversational assistant over the paper series and public track record. Audit-by-construction.

See tools in depth →

Contact