Job Description
acuCyber is seeking an AI Engineer to design and deploy secure AI systems that connect foundation models with enterprise tools, data, and services. You will build agent architectures, advanced retrieval capabilities, and stateful workflows that perform complex tasks while enforcing user permissions and maintaining reliable, controlled execution.
We are looking for a hands-on engineer who understands both the capabilities of autonomous agents and the discipline required to operate them in production. Strong candidates will combine AI orchestration and distributed systems experience with a practical approach to security, observability, and human oversight.
This is a full-time, on-site role at a Government facility in Charleston, SC or the National Capital Region (NCR).
Key Responsibilities:
• Agentic Architecture & Protocol Integration: Architect modular, interoperable agent ecosystems utilizing the Model Context Protocol (MCP) to standardize connections between foundation models, external tools, and enterprise data services. Design complex stateful workflows using production orchestration frameworks (LangGraph, CrewAI, Claude Agent SDK).
• Next-Gen RAG & Stateful Reasoning: Engineer advanced retrieval systems including Graph-based RAG, Agentic Routing, and Corrective RAG (CRAG). Design memory architectures that optimize context window utilization and reliably manage state across long-lived, multi-turn agent sessions.
• Data Governance & Entitlement Enforcement: Establish strict permission boundaries around autonomous tool calling. Map strict data serialization schemas (JSON Schema, Protocol Buffers) directly to enterprise Access Control Lists (ACLs) and RBAC/ABAC policies so agents only retrieve data and execute actions explicitly authorized for the invoking end-user.
• Runtime Guardrails & System Stability: Build resilient execution harnesses that prevent runaway agent loops and cascading failures. Implement idempotent tool retries, strict token and cost budgets, deterministic early-exit stop conditions, and seamless Human-in-the-Loop (HITL) approval gates for high-consequence actions
Required Skills
• Bachelor's degree in Cybersecurity, Computer, Electrical, or Electronics Engineering, or Mathematics with a concentration in computer science or equivalent
• Orchestration & Protocols: Deep, hands-on production experience with Model Context Protocol (MCP) server/client implementations, LangGraph, CrewAI, and the Claude Agent SDK. Proven ability to design custom, non-standard cognitive architectures when off-the-shelf patterns fall short.
• Advanced Retrieval & Context: Demonstrated expertise deploying GraphRAG (knowledge graphs + vector search), dynamic query routing, and self-correcting retrieval pipelines. Deep understanding of context compression, state serialization, and session persistence.
• Security & Schema Mapping: Strong proficiency with JSON Schema and Protocol Buffers (Protobuf). Proven track record of integrating identity and access management (IAM) tokens/claims into AI tool execution layers to enforce zero-trust, per-user data entitlements.
• Production Reliability: Practical engineering experience building fault-tolerant distributed systems, including idempotency keys for external API/tool mutations, implementing early-exit stop conditions, automated budget enforcement, and asynchronous HITL state suspension/resumption.
• 5+ years of relevant professional experience
• Willingness to work onsite at a Government facility full time.
• Active Top Secret/Sensitive Compartmented Information (TS/SCI) clearance with ability to obtain a Counterintelligence (CI) Polygraph.
Desired Skills
• Master's degree in Cybersecurity, Computer, Electrical, or Electronics Engineering, or Mathematics with a concentration in computer science or equivalent
• Evaluation & Observability: Experience implementing agent trajectory evaluation, step-level tracing (e.g., OpenTelemetry, LangSmith, Arize Phoenix), and automated red-teaming for prompt injection and privilege escalation.
• High-Compliance Environments: Background deploying AI systems within high-security, regulated, or mission-critical enterprise environments where auditability and data lineage are mandatory.
• Systems Programming: Proficiency in Python and TypeScript/Go for building high-throughput, low-latency MCP servers and custom orchestration runtimes.
Additional Details
Why Join acuCyber:
• Competitive compensation and performance incentives
• Full benefits package: health, dental, vision, 401(k) with match
• Paid time off and short-term disability coverage
• Purpose-driven, clearance-respecting culture
• Opportunities for professional growth and leadership
