Senior AI Engineer

The Opportunity

At Rohirrim, we're building the AI that powers how the government buys - from billion-dollar weapons systems to the software that keeps agencies running. Acquisition is slow, paper-heavy, and consequential; we're using agentic AI to make it faster and more defensible without lowering the bar on rigor. Your work will ship into production environments where the decisions really matter.

We're looking for a talented, self-driven Agentic AI Engineer to join our core engineering team and help build the next generation of autonomous AI systems that power our acquisition platform. You will design, develop, and deploy sophisticated multi-step AI agents capable of reasoning over complex procurement documents, managing long-horizon workflows, and delivering precise, defensible outputs in high-stakes environments.

This is a hands-on individual contributor role. You will work closely with our ML researchers, full-stack engineers, and product teams to ship agentic features that directly impact how government agencies and enterprise organizations acquire capabilities. Your work will be used in production environments handling sensitive and consequential decisions, so rigor, reliability, and security-consciousness are as important as innovation.

Note: This is a Hybrid Role - 3 days per week in our office in McLean, VA

Agent Design & Development

• Architect and build multi-agent systems and autonomous workflows for document analysis, requirement extraction, RFP response generation, and procurement pipeline automation.
• Design and implement agentic orchestration using frameworks such as LangGraph, AutoGen, CrewAI, or custom-built solutions tailored to Rohirrim's platform requirements.
• Develop tool-using, reasoning, and planning capabilities that enable agents to decompose complex acquisition problems into executable subtasks.
• Build robust human-in-the-loop and approval gates for high-stakes decision points within automated workflows.

Graph RAG & Document Intelligence

• Design and optimize Retrieval-Augmented Generation (RAG) pipelines to ground agent outputs in authoritative procurement data, regulations (FAR/DFARS), past performance records, and institutional knowledge.
• Implement advanced retrieval strategies including hybrid search, re-ranking, and context-aware chunking across large, heterogeneous document corpora.
• Develop document parsing and structuring pipelines for RFPs, SOWs, PWS, solicitations, and other complex government acquisition artifacts.

LLM Integration & Evaluation

• Select and integrate appropriate LLMs (commercial and open-source) for specific agent tasks, balancing capability, latency, cost, and security requirements.
• Build evaluation frameworks to systematically test agent accuracy, hallucination rates, and task completion against procurement-specific benchmarks.
• Implement prompt engineering best practices, including structured outputs, chain-of-thought reasoning, and few-shot prompting optimized for acquisition workflows.

Platform & Infrastructure

• Collaborate with platform engineers to deploy agents in production with observability, logging, tracing, and graceful failure handling.
• Contribute to shared tooling, internal SDKs, and reusable agent components that accelerate development across the engineering team.
• Participate in architecture reviews, code reviews, and technical planning sessions with a focus on scalability and maintainability.

What You'll Bring

Required

• 8+ years of software engineering experience, with at least 1 year focused on LLM applications, AI agents, or applied ML systems in production.
• Strong Python proficiency; experience building production-grade AI/ML pipelines.
• Hands-on experience with LLM orchestration frameworks (LangChain, LangGraph, AutoGen or equivalent).
• Deep understanding of RAG architectures, vector databases (Pinecone, Weaviate, pgvector or similar), and embedding models.
• Experience integrating APIs from frontier model providers and tuning prompts for structured and reliable results
• Strong software engineering fundamentals: clean architecture, testing, observability, and version control best practices.

Plusses

• Experience building AI products for regulated enterprise environments.
• Familiarity with federal acquisition regulations (FAR, DFARS) or proposal management processes.
• Experience with fine-tuning or instruction-tuning open-source LLMs (LLaMA, Mistral, etc.).
• Knowledge of secure-by-design principles and air-gapped or FedRAMP-compliant deployment environments.
• Active Secret or TS/SCI security clearance (or ability and eligibility to obtain).
• Experience with cloud infrastructure on AWS, Azure, or GCP; containerization with Docker/Kubernetes.

Success Profile

• A collaborative IC that is excellent at giving and receiving technical feedback, communicating trade-offs clearly, and elevating teammates.
• Ability to help define how agentic AI gets implemented in a SaaS company