Operations Research Engineer

About Pelico Traditional industrial manufacturing systems were built for predictability. Supply chains no longer are. Today, real-time data is buried in legacy tools that weren't designed for constant disruption. By the time an issue surfaces, it has already snowballed, sending everyone into firefighting mode. Pelico exists to close the gap between production plans and the daily reality on the factory floor. We've built the manufacturing orchestration platform trusted by the world's most complex industrial operations : companies like Airbus, Boeing, Safran, Eaton, and Cartier. Our AI connects fragmented supply chain, production, and quality data into a single live operational picture — then puts AI agents to work anticipating disruptions, simulating recovery options, and keeping every team aligned before small problems become expensive ones. Founded in 2019 and backed by General Catalyst, 83North, and Serena, we've raised $72M and are growing fast — 300% year-over-year revenue growth, deployed in 15+ countries, and supporting factories worldwide. We're headquartered in Paris, with offices in Miami, London and Amsterdam, and we're just getting started. This is a rare chance to join a company that's redefining an industry that hasn't been disrupted in decades — from the inside. About the Role Manufacturers run on systems of record — ERP, MES. Pelico sits on top and turns that raw data into a live, realistic model of the plant: a supply-chain twin that projects what will actually be produced and shipped, and where it will fall short. Pelico already goes further than flagging risk — it prescribes actions and lets planners act on them, writing decisions back to the systems of record. Pelico is already both a system of information and a system of action. The next frontier is to take that much further. In a real factory, exceptions and disruptions never stop — a late part, a quality hold, a slipped operation — and the plan drifts from execution within hours. We want every decision-maker — human planners and, increasingly, AI agents — resolving those disruptions continuously (expediting or chasing a supplier, pulling in / pushing out orders, re-allocating scarce stock, re-sequencing work orders, sourcing an alternative), all pulling in the same direction: the next best action prescribed for each, every one of them unified against a single objective. That direction is a single, configurable North Star (OTD, turnaround, inventory €, revenue). The twin is both the world every action plays out in and the judge that scores it — pricing any action’s global impact so local decisions converge to the company objective instead of fighting each other. It’s a hard, open problem: multi-objective, continuous and reactive, bounded by real feasibility, trustable and explainable, and increasingly multi-agent — agents in conflict, order-dependent outcomes, arbitration. Expect the full toolbox: exact solvers where they fit, neighborhood search and metaheuristics where they don’t, and reinforcement-learning-style formulations where the twin is the environment and the North Star is the reward. What You'll Do Own the optimization strategy — define the vision and multi-quarter roadmap for optimization at Pelico: what to build in-house vs. buy, what to sequence when, which bets to make, each tied to the company’s objectives. You’re accountable to leadership for the outcomes and set the technical direction the whole discipline follows. As one of the company’s senior technical voices, you have a seat at the table, helping shape product and company strategy, not only your own function. Build and lead the team — found, hire, and grow a credible optimization team from the ground up: develop careers, set the technical and cultural bar, and own its performance and delivery. You are the draw that pulls top OR talent toward a company not yet known for it. Partner with product and engineering leadership to embed optimization across the roadmap. Represent Pelico’s OR to the world — publish, speak, and represent Pelico in the field to build our research and employer brand. Forge partnerships with universities and research labs that feed both the science and the hiring funnel. Back the commercial front — be an asset in strategic deals, where trust is earned, not assumed. Lend pre-sales credibility, win over skeptical customers, and build the senior technical relationships that turn optimization into revenue. Drive the core optimization — allocation and recommendations beyond FIFO (multi-objective, under real constraints); the North Star scoring and conflict-arbitration layer that aligns humans and agents; optimization under uncertainty (lead/cycle times and yield as distributions, not point estimates); and stock/safety-policy optimization computed from our own historical actuals, not declared ERP parameters. Own trust and adoption — an optimizer only creates value when a planner acts on it. Own the optimality-vs-runtime trade-off deliberately, keep every recommendation explainable, and validate it against how plants really behave. Explainable and trusted is a hard constraint here, not a nice-to-have. Enable agent-actionable optimization — expose the twin and its scoreboard as clean, governed, explainable tools an AI agent can call to read the plan, simulate an action, score it, and act under human-in-the-loop gates. Lead by example — author the models and specs, write and review code, and mentor strong engineers into optimization technique. At this level, the leadership and the engineering are the same job. About You A PhD in Operations Research, Applied Mathematics, or a related field, and a senior track record solving real optimization problems in industry. Recognized thought leadership — a body of work (publications, patents, talks, open-source) and the authority that comes from having cracked genuinely hard problems. Deep production experience across combinatorial optimization (MILP/LP/CP), metaheuristics, and simulation — equally comfortable reaching for an off-the-shelf solver (Gurobi / CPLEX / OR-Tools / Hexaly) and hand-crafting a bespoke algorithm when no solver fits or is fast enough, at scale (millions to hundreds of millions of records). The judgment to make the exact-vs-heuristic, optimality-vs-speed trade-off deliberately — and the drive to build a team and a discipline, not just models. At home in an AI-native environment — comfortable packaging optimization as tools that AI agents can consume. Still hands-on in the code: the strongest in the room, not the one who left the keyboard. Bonus: supply-chain / manufacturing planning depth (MRP, multi-level BOM, pegging, scheduling); optimization packaged for agent consumption. Pelico is an equal opportunity employer. We're committed to building a diverse team and an inclusive environment where everyone can do their best work.