Staff Data Scientist, Optimization

Who We Are At Firestorm, we are building the future of expeditionary defense manufacturing and autonomous systems. Modern conflict has exposed a fundamental problem: the systems needed most by operators are often too expensive, too slow to produce, and too difficult to sustain at scale. Firestorm exists to change that. We develop mission-adaptable aerial systems and deployable manufacturing infrastructure designed to put capability directly into the hands of the warfighter. From modular unmanned aircraft to xCell — our deployable microfactory — our goal is to make defense systems rapidly deployable, adaptable, and producible at the point of need. We are looking for builders, operators, and problem-solvers who want to work on meaningful technology with real-world impact. About the Role Firestorm builds uncrewed aircraft and the manufacturing network that produces them: full-capability plants, deployable xCell edge factories, and forward-deployed sites that stand up production near the point of need. Crucible is the software that runs that network. It plans and executes production across factories that don't look alike, using supply that isn't guaranteed to arrive. It has to answer practical questions: what should we build, where should we build it, what constrains the plan, and what changes when material, capacity, or demand moves. Many of those are optimization problems. [... source excerpt omitted ...] it produced the plan it did. We are hiring a Staff Data Scientist to own the optimization and decision science behind Crucible. As the first dedicated data scientist on the team, you'll set the technical direction for how Crucible models and solves planning problems, while staying hands-on building the models that ship in the product. You'll work directly with the Crucible software team and with manufacturing, supply chain, and product as we grow both the platform and the science behind it. What You’ll Do Own Crucible's optimization models — Formulate and solve problems across production planning, allocation, sourcing, inventory, scheduling, and manufacturing network optimiz [... source excerpt omitted ...] h — Use mathematical optimization, heuristics, simulation, statistics, or simpler methods based on the problem and the decisions the user needs to make. Build models that work in production — Prototype against real data, establish baselines, test solution quality and performance, and work with software engineers to ship models inside Crucible. Model real-world uncertainty — Account for incomplete data, changing demand, variable lead times, supplier performance, and other conditions that make a mathematically perfect plan operationally wrong. Work directly with users — Spend time with manufacturing, supply chain, and program teams to understand how decisions are actually made and