AI Security Engineer

Ampcus Inc.•Gaithersburg, MD

12h•Onsite

About The Position

Ampcus Inc. is a certified global provider of a broad range of Technology and Business consulting services. We are in search of a highly motivated candidate to join our talented Team. This role focuses on ensuring the security of AI systems, including supply chain security, AI attack simulation, and runtime security.

Requirements

Experience with Supply Chain Security/Model Scanner (SaaS self-hosted hybrid deployment)
Experience maintaining internal model scanning pipeline infrastructure
Experience monitoring container releases and updates
Experience reviewing model scanner output
Experience setting organizational policy around acceptable risk in model vulnerabilities
Experience with AI Attack Simulation (SaaS-only)
Experience maintaining AIAS tasks implemented into application pipelines
Experience adding AIAS tasks to new applications or running on-demand attack simulations
Experience reviewing attack simulation results
Experience setting organizational policy around acceptable risk in applications
Experience providing remediation guidance to application teams
Experience with Runtime Security/AIDR (self-hosted hybrid deployment)
Experience maintaining self-hosted HiddenLayer environment
Experience ensuring availability for applications via AI/LLM Gateway
Experience ensuring appropriate ingestion of runtime logs and results in SIEMs and SoRs
Experience setting organizational policy around runtime behavior
Experience determining alerting thresholds and appropriate warning or escalation behavior
Experience reviewing alerts and escalations from HiddenLayer
Experience ensuring new applications are appropriately protected by HL Runtime Security

Responsibilities

Maintain AZ's internal model scanning pipeline infrastructure and ensure continued appropriate ingestion in compliance and governance systems (in the case of self-hosted deployment).
Monitor container releases and updates for the model scanning pipeline.
Review output from the model scanner to ensure it is meeting requirements.
Set organizational policy around "acceptable" risk in model vulnerabilities and provide internal guidance and overrides where necessary.
Maintain AIAS tasks that are implemented into application pipelines.
Add AIAS tasks to new applications or run on-demand attack simulations.
Review attack simulation results and determine application risk levels.
Set organizational policy around "acceptable" risk in applications and provide remediation guidance to application teams.
Maintain the self-hosted HiddenLayer environment and ensure availability for AZ applications (via the proposed AI/LLM Gateway).
Ensure continued appropriate ingestion of runtime logs and results in SIEMs and SoRs.
Monitor container releases and updates for runtime security.
Set organizational policy around runtime behavior (blocking vs alerting) in applications.
Determine alerting thresholds and appropriate warning or escalation behavior for identified runtime misuse/attacks.
Review alerts and escalations from HiddenLayer and take appropriate action (remediation, tuning policy for specific applications, etc.).
Ensure that new applications are appropriately protected by HL Runtime Security (via the proposed AI/LLM Gateway).