Adversarial AI Software Engineer (mid-career)

Lockheed Martin•Dallas, CO

1d•Hybrid

About The Position

We are seeking an experienced software engineer to support a small team to design, develop, integrate, and build an AI red teaming platform. This role is focused on developing advanced adversarial AI tools and capabilities in support of the Lockheed AI Center (LAIC) Red Team operating across classified and unclassified environments. This position integrates open source, custom developed, and commercial adversarial AI utilities, provides a unified user interface, and enables automated, repeatable assessments of AI systems across use cases, including: generative/agentic, computer vision, and RF/EW. These AI systems span the deployment spectrum from Lockheed Martin internal enterprise and cloud environments to edge compute platforms. The position involves both planned project work and rapid ad hoc support, with extensive documentation, automation, and occasional travel for field demonstrations.

Requirements

Demonstrated experience in software development in cloud or Kubernetes-cluster environments.
Hands-on engineering experience with building tools, frameworks, and automation integrated with GitLab, cloud, DevSecOps, or CI/CD pipelines.
Demonstrated experience developing microservices (preference for Python).
Demonstrated ability to author high‑quality documentation (design specs, runbooks, API guides).

Nice To Haves

Demonstrated experience in systems engineering, software architecture and design, test-driven development, and collaborative reviews.
Experience containerizing applications (i.e., Docker, Podman) and deploying them to scaled Kubernetes/OpenShift environments via infrastructure-as-code (e.g., Ansible, Helm).
Experience building AI agents using frameworks (e.g., CrewAI, AutoGen, LangChain/LangGraph).
Experience working with vector databases (e.g., AWS OpenSearch, Weaviate).
Experience with modern web frameworks (e.g., React, NextJS).
Experience working with deep learning frameworks (e.g., PyTorch, TensorFlow).
Experience with the Lockheed Martin AI Factory platforms (i.e., Panel, Genesis, Navigator).
Experience with adversarial machine-learning, AI security research, and implementation.
Experience administering Splunk and dashboards/alerts with Prometheus & Grafana.
Experience using LangFuse and OpenTelemetry for traceability and observability.
Experience with virtualization/hosting platforms (VMware, Hyper V, KVM) and secure remote access workflows.
Experience applying secure-coding best practices, STIGs, vulnerability remediation, and automated CI/CD security testing.
Familiarity with GPU compute ecosystems (NVIDIA drivers, CUDA, container runtimes) and performance profiling.
Ability to work effectively across classified and unclassified environments, adhering to all applicable security policies.
Familiarity with AI security or adversarial ML.
Current DoD 8570 baseline certification (IAT II or higher).

Responsibilities

Collaborate closely with multiple AI Red Teams to design, architect, and implement adversarial AI capabilities for autonomous agents, computer vision systems, and RF and EW systems, translating state‑of‑the‑art research into production‑ready code.
Design and build a modular, extensible red‑teaming framework that enables repeatable, user‑driven evaluations.
Design user-facing interfaces and workflows (CLI, web UI, APIs, CI/CD pipelines) and documentation to enable internal and future third-party users to execute red team activities with minimal subject matter expertise.
Integrate open‑source, commercial, and inner-source solutions into a cohesive platform, handling packaging, versioning, release, dependency management, and maintenance.
Develop and maintain CI/CD pipelines (GitLab, Azure DevOps, or equivalent) with automated testing, security scans, and deployment to cloud and containerized/VM‑based environments.
Coordinate with Infrastructure Operations (InfraOps) stakeholders for virtualized, containerized, and cloud-based Linux environments, including configuration, orchestration, DISA STIG compliance, and lifecycle maintenance.
Implement observability (structured logging, metrics, dashboards) using Splunk, Prometheus, Grafana, LangFuse, OpenTelemetry, and ensure full traceability for debugging, performance analysis, operational reliability, and audit.
Develop GPU-enabled software workloads (NVIDIA drivers, CUDA, container runtimes) and continuously optimize performance and stability.
Champion secure development practices – embed threat modeling, code reviews, and DevSecOps controls throughout the software lifecycle.
Coordinate development and deployment activities across classified (on-site) and unclassified (remote) environments while ensuring secure and reliable operations.