Job Summary

Own the agentic systems at the center of Orbis AI — the orchestration, planning, retrieval, and tool-use machinery behind the Agent Catalog and the OSINT Agent Harness. You will build agents that hold up under real mission workloads and set the evaluation bar the rest of the organization writes against. This is senior engineering on AI systems: the hard problems are correctness, explainability, cost, and behavior at scale.

Travel: Up to 15% for customer engagement, integration support, and team collaboration.

Key Responsibilities

• Architect and ship agents across the portfolio — knowledge agents over customer data, tool agents that wrap external operations, and goal-driven agents that plan and execute multi-step analysis.
• Own the orchestration layer: planning, tool selection, state and memory, retries, partial failure, and termination behavior.
• Design and own retrieval end to end — chunking, embedding, indexing, ranking, and the query-time behavior that decides answer quality.
• Build the provenance and citation paths that let a user see how a conclusion was reached, and treat unexplainable output as a defect rather than a limitation.
• Own the evaluation infrastructure for agent quality — accuracy, groundedness, explainability, latency, and cost — and set the standard other engineers write cases against.
• Integrate model providers and third-party tools behind stable internal interfaces, and instrument agents so quality regressions and cost anomalies surface before a customer finds them.
• Partner with research engineers to productionize validated results, and mentor engineers moving into AI work from a conventional software background.
• Travel up to 15% for customer engagement, integration support, and team collaboration.

Required Qualifications

• 4+ years of professional software engineering experience, including 2+ years building LLM-backed or agentic systems in production
• Bachelor's degree in Computer Science, a related technical field, or equivalent professional experience
• Strong programming skills in Python and/or TypeScript, with substantial production code in at least one
• Deep practical experience with retrieval-augmented generation — embeddings, vector search, chunking strategy, ranking, and the failure modes of each
• Demonstrated experience building and debugging multi-step, tool-using agents, whether on a framework or from primitives
• Proven rigor about evaluation, including harnesses that measured whether an AI change actually improved anything
• Willingness to travel up to 15%

Desired Qualifications

• Experience with knowledge graphs, ontologies, entity resolution, or information extraction
• Experience running open-weight models in constrained or disconnected environments where a hosted API is not an option
• Experience with production observability and cost control for model-backed workloads
• Familiarity with OSINT, intelligence, or research workflows and the accuracy and provenance expectations that come with them
• Experience mentoring engineers into AI work from a conventional software background
• Open-source contributions, publications, or academic collaboration in applied AI

Physical Requirements

• Prolonged periods of sitting at a desk and working on a computer
• Participation in virtual and in-person meetings
• Ability to attend planned meetings and/or work in classified spaces for extended periods within the specified work regions
• Travel: Up to 15% for customer engagement, integration support, and team collaboration.