Radu Ionescu, ai security engineer

Radu Ionescu

Senior AI Security Engineer

Cluj-Napoca, Romania · Europe/Bucharest (UTC+3) · 11 years of experience

Radu focuses on securing AI applications, data paths, model supply chains, and agent permissions.

Request a shortlist

Why this profile fits

AI-specific threat modeling

Covers prompt injection, sensitive-data exposure, unsafe tool use, model supply chains, and excessive agency.

Least-privilege architecture

Scopes identities, tools, data access, secrets, and network paths to the minimum required for each workload.

Adversarial validation

Turns abuse cases and incidents into repeatable tests, monitoring signals, and response playbooks.

Relevant toolkit

Threat ModelingIAMKubernetes SecurityLLM Red TeamingSIEM

Delivery evidence

Selected work

GPU Workload Security Baseline

cloud platforms

Designed and delivered a AI security programme focused on Kubernetes security for AI workloads; reduced critical cluster-policy findings by 71% across GPU namespaces. The release included threat models, least-privilege controls, adversarial tests, logging, and incident playbooks mapped to OWASP guidance.

Threat ModelingPython

Kubernetes Policy Verification Pipeline

cloud platforms operations

Built the supporting control and measurement layer for the primary system, covering review queues, regression checks, operational visibility, and a documented handoff to the owning team.

Threat ModelingKubernetes Security
View more profile detail →

Relevant experience

Career timeline

Senior AI Security Engineer

2023–Present

Devlyn client assignments · Remote

  • Led Kubernetes security for AI workloads delivery for a cloud platforms team and reduced critical cluster-policy findings by 71% across GPU namespaces.

AI Security Engineer

2015–2022

cloud platforms product company (confidential) · Cluj-Napoca, Romania

  • Built production systems in cloud platforms, with increasing ownership of reliability, testing, and stakeholder delivery.

Skill depth

Show experience in context.

VerbalCommunicationDomainUnderstandingProblemSolvingCodeQualitySystemDesignDeliverySpeed
Competency shapeAssessment across the same six dimensions used for every profile.

Relevant experience by skill

Threat Modeling11 years
Python10 years
IAM9 years
Kubernetes Security8 years
LLM Red Teaming11 years
SIEM10 years
See the complete skill index

Languages

PythonTypeScriptSQLBash

Frameworks

PythonIAMKubernetes SecurityOPAVaultSemgrep

Tools

Burp SuiteFalcoTerraformGitHub ActionsDocker

Platforms

AWSGoogle CloudAzureKubernetes

Storage

VaultS3PostgreSQLSecurity Lake

Paradigms

Secure-by-design AIZero trustAdversarial testingSupply-chain security

Credentials

Certifications

Certified Cloud Security Professional

ISC2 · 2026 · Certified

HashiCorp Certified: Vault Associate (003)

HashiCorp · 2025 · Certified

Communication

Languages

English

Fluent

More examples

Similar AI Security Engineer profiles.