AI Engineer – Decision & Optimization Systems

AI Engineer – Decision & Optimization Systems About Gallatin

At Gallatin, we are rebuilding defense logistics for the warfighters of the United States and allied forces. We take an AI-first approach to modernizing how materiel, fuel, and equipment move from factory to foxhole—so commanders and operators can make faster, better, and safer decisions under uncertainty.

We believe that transformational systems are built shoulder-to-shoulder, not over Zoom, and we operate from our offices in El Segundo, California and Austin, Texas.

We’re looking for an AI Engineer – Decision & Optimization Systems to own the feasibility, authority, and constraint layers that sit between AI reasoning and real-world execution.

In this role, you will build the systems that decide what plans are allowed to exist before any optimizer, agent, or human can act on them. Your work ensures that AI-generated logistics plans respect command hierarchy, policy, safety, and operational reality—so when we execute, we execute correctly.

You will work closely with routing, packing, optimization, and AI-agent teams to translate intent, authority, and uncertainty into auditable, enforceable, and explainable decision systems.

Build & Own Feasibility and Authority Models

• Design and implement models that encode command hierarchy, authority limits, and policy rules into formal feasibility checks.

• Convert AI-agent outputs (LLMs, planners, probabilistic models) into deterministic, enforceable constraints before execution.

• Build and maintain a centralized constraint framework used across optimization and planning systems.

• Encode:

  • Authority rules

  • Compatibility constraints

  • Timing windows

  • Asset availability and readiness

• Surface clear diagnostics for infeasible plans and constraint violations so humans and machines can understand what failed and why.

• Own the data inputs required for feasibility and constraint enforcement.

• Define and enforce data contracts, normalization, and validation pipelines.

• Build robustness against incomplete, delayed, or noisy operational data.

• Integrate feasibility checks into real-time and batch planning pipelines.

• Partner with optimization and execution teams to gate all actions on valid, policy-compliant inputs.

• Validate system behavior using scenario testing, simulations, and operational feedback from real users.

Strong Engineering & Modeling Skills

• Proficiency in at least one of:
  Python, Java, Go, or C++.

• Experience building constraint-based or rule-based systems in production.

• Experience integrating AI or agent-based reasoning into downstream decision or execution pipelines.

• Degree or equivalent experience in Operations Research, Applied Math, Industrial Engineering, Computer Science, or related field.

• Strong understanding of:

  Linear programming

  Mixed-integer programming

  Constraint satisfaction systems

• Experience using tools such as Gurobi, CPLEX, OR-Tools, Pyomo, AMPL, or equivalent.

• Ability to reason about feasibility vs optimality tradeoffs.

• Comfort integrating probabilistic or AI‑derived outputs into deterministic systems.

• Strong instincts for debugging, failure analysis, and explainability in high‑stakes environments.

• Experience with logistics, supply chain, or routing systems.

• Experience working alongside optimization or planning teams.

• Exposure to defense, government, or mission‑critical operations.

• Prior work integrating AI agents into decision or execution pipelines.

• Join a mission‑driven, high‑impact startup building the decision systems that power modern defense logistics.

• Work on real operational problems where correctness, safety, and explainability matter.

• Collaborate with world‑class engineers, operators, and AI researchers.

• Competitive compensation including generous equity, 100% employer‑paid health insurance, 401(k), unlimited PTO, and catered meals.

This position requires the ability to obtain and maintain a U.S. security clearance and to work in classified environments when necessary. If you’re not excited about building AI systems that decide what is allowed to happen in war, this role is not for you.