DoD AI/ML Engineer — JWICS & Cross-Domain

ECS Corporate Services; Washington (District of Columbia); On-site; USD 125,000 - 145,000; Full time

Benefits offered by this job

Security clearance support

Job summary

Everforth ECS is seeking an AI/ML Engineer to work onsite at the Pentagon in Arlington, VA, supporting the PSP2 Agentic Network for a DoD customer. The role involves designing, deploying, and sustaining secure agentic AI, data, and platform capabilities across classified and unclassified environments, with emphasis on JWICS-hosted services.

The candidate will evaluate hosting architectures, implement cross-domain data transfer patterns, and develop observable, maintainable systems.

Qualifications

• Bachelor’s degree in Computer Science, Information Technology, Software Engineering, Systems Engineering, Cybersecurity, or a related technical field is required. A master’s degree is preferred.
• 5+ years of relevant professional experience supporting technical integration, deployment, test, fielding, or sustainment activities.
• Ability to work onsite at the Pentagon in Arlington, VA.
• Ability to travel < 50% to CONUS/OCONUS customer sites and government installations.

Responsibilities

• Manage, supervise, and lead the technical architecture for the PSP2 Agentic Network, including agent orchestration, model inference, data access, APIs, platform services, network boundaries, observability, and secure deployment patterns.
• Assess the technical suitability, performance, resilience, support model, and operational constraints of JWICS hosting options and document patterns and tradeoffs.
• Deploy and operate LLM inference and agentic AI services, including model-serving endpoints, prompt and policy controls, retrieval and tool-use integrations, scaling, performance tuning, monitoring, and failure recovery.
• Implement a dedicated unclassified platform for controlled storage, discovery, access, and sharing of releasable data with approved allies and partners.
• Analyze data flows between unclassified and JWICS environments; define data prep, labeling, format, security, release, and interface requirements for cross-domain movement.
• Develop technical designs, integration specs, and proofs-of-concept for cross-domain data-transfer patterns without bypassing security or authorization processes.
• Implement observability, health checks, logging, metrics, tracing, alerting, and operational dashboards.
• Perform technical hardening, testing, security validation, and operational-readiness verification before release to users.
• Troubleshoot complex issues in environment config, software dependencies, network connectivity, IAM, interfaces, containers, data pipelines, and deployment processes.
• Produce and maintain architecture diagrams, interface-control docs, baselines, runbooks, troubleshooting procedures, deployment instructions, and technical decision records.
• Evaluate architectural alternatives and articulate tradeoffs involving security, reliability, performance, scalability, interoperability, maintainability, and cost.
• Provide artifacts to support cybersecurity, accreditation, and Authorization to Operate activities.
• Review implementations, set engineering standards, and guide developers working on PSP2 Agentic Network.
• Provide on-site and remote support during integration, deployment, and stabilization periods.

Skills

AI/ML engineering

Security architecture

JWICS hosting

Cross-domain solutions

Observability

Troubleshooting complex systems

Education

Bachelor’s degree in CS/related field

Master’s degree preferred

Tools

AWS Cloud

Security+

Job description

Everforth ECS is seeking an AI/ML Engineer to work onsite at the Pentagon in Arlington, VA, supporting the PSP2 Agentic Network for a DoD customer. The role involves designing, deploying, and sustaining secure agentic AI, data, and platform capabilities across classified and unclassified environments, with emphasis on JWICS-hosted services.

The candidate will evaluate hosting architectures, implement cross-domain data transfer patterns, and develop observable, maintainable systems.