DIRECTOR OF AGENTIC SYSTEMS
Fully Remote | U.S. Person Required | Immediate Start
Reports To: Director of Technology
Level: Director
Clearance: U.S. Person required; Secret-eligible preferred
ABOUT THE OPPORTUNITY
We operate a production fleet of autonomous AI agents — each with a defined identity, a delegated authority, a credential vault, and a dedicated project workspace. The fleet covers functions including business development, capture management, SBIR program leadership, financial control, technology direction, engineering, operations analysis, organizational psychology, executive coordination, and secretariat support. We are looking for the person who makes the fleet run.
We are a defense electronics holding company. The portfolio spans avionics MRO, government contracting across DoD, USAF, USSF, and MDA, SBIR/STTR R&D programs in hypersonic thermal protection and advanced sensors, and an active M&A pipeline. Three operating entities run under a unified AI-first governance model with formal delegation and mathematical authority proofs.
We are building a sovereign, composable, AI-first operating system: our servers, our models, our stack. The target state has zero hyperscaler dependency.
THE ROLE
Director of Agentic Systems is the delivery owner for everything that makes the autonomous agents function at scale.
You are not the person deciding how models reason — that responsibility sits with the Director of AI Engineering. You decide how the system operates: which workflows fire, which pipelines move data, which integrations connect operational systems to the agent layer, and which coupling points pass work between agents without anything getting dropped.
If the agent fleet is a vehicle, the AI Engineering team tunes the engine. You build the drivetrain, the wiring, the fuel system, and the dashboard — and you make sure that when the accelerator goes down, the vehicle moves.
This is an architect-builder role. You design the system and you build it. The agent fleet itself handles your documentation, status reports, and coordination, so your time stays in the work only you can do.
WHAT YOU WILL OWN
1) Agent Fleet Delivery Layer
Our autonomous agent runtime platform is yours to configure for production: agent personas, tool registrations, memory integration, runtime parameters, and the handoff protocol that moves work cleanly between agents. When an opportunity surfaces in SAM.gov, gets scored by our intelligence platform, and routes to capture development, you built the system that makes that handoff happen.
2) N8N Workflow Engineering
Our operational data does not move itself. DIBBS solicitations arrive by email and need to be triaged, scored, and routed. Repair orders have lifecycle stages that trigger invoicing events in Business Central. SBIR milestones generate billing obligations. International BD activities create follow-up workflows across three time zones. You build production N8N flows — not demos — with error handling, audit trails, and retry logic.
3) Government Intelligence Feed Into the Agent Layer
Our internal government intelligence platform ingests SAM.gov, scores every opportunity against a multidimensional evidence model, and validates claims through a formal truth protocol with provenance hashing. You own how that intelligence reaches each agent who needs it, and you own the sovereign retrieval node that injects relevant past context into every agent reasoning loop.
4) Fleet Expansion and M&A Onboarding
The fleet is scaling from 10 agents to 29+. You will design and execute the per-agent provisioning pipeline and establish a terminal-to-human adjudication gate as the standard pattern for intelligence-to-action workflows. When acquisitions close, you build the automation that takes a new entity from Day 0 to fully agent-provisioned and integrated within 60 days.
WHAT WE ARE LOOKING FOR
7+ years building real integrations and workflow systems in production — systems with real users, real data, and real consequences when they break. Production systems, not throwaway prototypes.
Hands-on experience with:
- N8N, Make, Zapier, or equivalent at production scale; complex multi-step flows with error handling and monitoring
- REST API integration across disparate enterprise systems; you know what breaks and why
- Python: async patterns, queuing, event-driven architectures
- Process modeling in BPMN 2.0 or equivalent; you can read and write formal workflow specifications
- Enterprise data flows: ERP integration and operational data into AI layers
- U.S. Person status (required; our workflows touch ITAR-controlled defense data)
Additional helpful experience:
- Multi-agent AI frameworks: LangChain, CrewAI, AutoGen, Claude Agent SDK, or equivalent
- Kubernetes / K3s; agent workloads deploy via GitOps into sovereign K3s
- Microsoft Dynamics 365 / Business Central integration (ERP backbone across all entities)
- Government contracting data: SAM.gov API, DIBBS/DLA intake, SBIR reporting
- Azure DevOps API depth (universal agent backbone)
- Active Secret clearance, or the willingness and ability to obtain one
WHY THIS ROLE IS UNUSUAL
You will build something that has not existed before. An autonomous agent fleet operating a multi-entity defense electronics company is not a reference architecture you can download. There is no playbook. You write it.
The scope is real from day one. Not a pilot. Not a sandbox. Production agent workflows running against live business data — government contracts, financial operations, engineering deliverables — for a defense contractor whose business outcomes depend on your work running correctly.
You operate with genuine authority. Director-level role, owning a defined Line of Effort in a formal operational program. Reports to the Director of Technology. You hire your own team as the fleet grows.
The company is built around this work. The Executive Director, the governance framework, and the operating model are all designed around the premise that autonomous agents and human leadership operate as peers. You are not an outlier — you are a core organizational function.