Research Engineer, AI

Orbis is a technology company that delivers sovereign intelligence and decision capabilities to the United States Government and its allies. We convert complex information and operational systems into decisive outcomes, giving public institutions and critical enterprises the tools to move faster, see clearer, and act first in an increasingly contested world.
We are not a consulting firm and not a systems integrator. Lead applied research in one or more Orbis AI focus areas — agentic AI, knowledge representation, dynamic ontologies, autonomous analysis, and high-performance analytics — and carry the results into the product. You will frame the problem, survey what already exists, build prototypes that test a real hypothesis, and measure honestly enough to act on the answer. The role exists because our advantage comes from turning advanced research into dependable mission capability.
Travel : Up to 20% for conferences, customer engagement, and team collaboration.
Frame research problems drawn from real product and mission gaps, define what a result must show to be worth productizing, and survey existing work before building.
Run rigorous experiments — baselines, ablations, held-out evaluation, and honest error analysis — with methodology another engineer can reproduce.
Transition validated results into production alongside product engineers, staying involved through the handoff rather than delivering a notebook.
Contribute to the scientific community through publications, open-source releases, and academic collaborations, and advise Orbis on what in the field is worth adopting.
Travel up to 20% for conferences, customer engagement, and team collaboration.
3+ years of applied research or research-engineering experience, or an M.D. in Computer Science, Machine Learning, or a related field with equivalent applied work
Strong Python engineering skills, producing research code another engineer can read, run, and build on
Depth in at least one Orbis AI focus area: agentic AI, knowledge representation, dynamic ontologies, autonomous analysis, or high-performance analytics
Clear technical writing that makes a result legible to engineers and non-specialist stakeholders alike
Willingness to travel up to 20%
Publication record, open-source contributions, or academic collaborations in a relevant area
Experience carrying a research result all the way into a production system
Experience with knowledge graphs, ontology engineering, reasoning systems, or information extraction
Experience with agent architectures, planning, or multi-agent coordination
Experience with model training, fine-tuning, or evaluation at scale
Experience with performance engineering on data-intensive or analytical workloads
Familiarity with OSINT, intelligence, or investigative research workflows
Prolonged periods of sitting at a desk and working on a computer
Participation in virtual and in-person meetings
Travel : Up to 20% for conferences, customer engagement, and team collaboration.
the chance to build technology that gives the United States Government and its allies a genuine decision advantage, not just another dashboard. From competitive retirement matching to a PTO policy that encourages real time off, were committed to creating an environment where our team can do their best work and still have a life outside of it.
Depending on the role, that might mean five days a week on a customer site, full-time in one of our offices, a hybrid schedule, or fully remote work; Travel is common across many Orbis roles, particularly those supporting customer missions in the field.
Orbis is headquartered in McLean, VA, with additional office locations in Taipei, Taiwan, and Canberra, Australia.
Orbis does not discriminate on the basis of race, color, religion, sex, sexual orientation, gender identity or expression, national origin, age, disability, protected veteran status, genetic information, or any other status protected by applicable law.
Reasonable accommodation is available to applicants and employees with disabilities upon request.