Researcher · Speaker · Core Framework Author

Architecting defensive telemetry, behavioral enforcement, and forensic audit infrastructure for autonomous agentic systems.

I specialize in identifying systemic, logic-based vulnerabilities in autonomous AI systems and building tamper-evident security frameworks for enterprise networks and agentic clouds. My focus is engineering crisp accountability vectors for non-human identities before they execute critical downstream transactions.

• OWASP AI Exchange: Core Framework Author contributing to global definitions for autonomous system safety and model trust boundaries.
• AIVSS v0.8 & OASec 2026: Named Specifications Reviewer and Founding Member Circle contributor mapping real-time multi-turn defensive baselines.
• Industry Speaker: Presenting advanced technical research on adversarial AI and non-human identity manipulation at Black Hat MEA 2025, DEF CON 33, and the upcoming ISC2 Security Congress 2026 (Denver, CO — "Architecting for Non-Human Identity in Agentic Clouds").
• Lead Architect @ ARTONexa: Engineering the ARTO Sovereign Security Suite (Tessera, VerityFlux, Vestigia) to deliver zero-trust IAM, trajectory-level behavioral enforcement, and hash-chained forensic logging.

• Agentic AI Security (AI Red Teaming): Mapping multi-turn session trajectory decay, prompt extraction, and logic manipulation across deep recursive tool-execution states.
• Non-Human Identity (NHI) Architecture: Formulating strict cryptographic delegation-chain tracking, scoped token lifecycle constraints, and immediate Redis-backed edge revocation mechanics.
• Forensic Audit Trails: Developing high-density, tamper-evident OpenTelemetry logging structures with SHA-256 validation chains for automated system verification.

• 📂 Execution Ecosystem   : Python (Expert), Rust (Systems Architecture), SQL, Go
• 🧠 Autonomous AI Core    : LangChain, LlamaIndex, PyTorch, HuggingFace Transformers
• 📡 Telemetry & SecOps     : OpenTelemetry Native, Redis, PostgreSQL, Docker, FastAPI
• 🛡️ Defensive Operations : Trajectory Tracking Engines, DPoP Token Binding, VAPT Scripting

• ARTONexa Suite — An interconnected behavioral enforcement stack featuring zero-trust IAM with depth-5 delegation logic, real-time multi-turn adversarial trajectory monitoring, and hash-chained database audit layers.
• ot-threat-intelligence-agent — An autonomous reasoning engine optimized for structural triage parsing, deep session trace diagnostics, and systemic threat evaluations.
• adversarial-agent-sim — A diagnostic framework built to evaluate boundary controls, privilege escalation risks, and data logic persistence errors within multi-agent cloud loops.

• Enterprise Integration Routing: miracle@artonexa.com
• Technical Deep-Dives: arksher.substack.com
• Professional Ledger: LinkedIn Profile
• Research Workspace: Research Discord

"Security is not a product, but a process of constant adversarial reasoning."