Julia MTK Modeling Engineer

📋 Description

• • Architect high-fidelity, acausal physics simulators in Julia and the ModelingToolkit (MTK) ecosystem that become the "textbooks" from which Archie learns how chillers, cooling towers, air handlers, hydronic networks, switchboards, transformers, and distribution systems behave under every conceivable operating point, fault mode, and design variant.
• • Own the end-to-end pipeline: white-board the governing equations, encode them into MTK components, validate against empirical data, orchestrate cloud-scale parameter sweeps that generate millions of synthetic traces, and package the datasets so that downstream ML engineers can train Archie to reason quantitatively about real-world physical systems.
• • Push the boundaries of numerical performance—profile sparse Jacobians, exploit symbolic tearing, parallelize DAE solves across thousands of CPU/GPU cores, and invent novel model-order reduction techniques so that a single simulation run drops from minutes to seconds without sacrificing the thermodynamic accuracy that determines Archie’s predictive power.
• • Collaborate in tight, cross-functional sprints with deep-learning researchers who turn your simulation outputs into tokenized training corpora; iterate daily on model fidelity vs. compute cost trade-offs, feeding real-world telemetry back into the simulator to close the loop between synthetic data and on-site performance.
• • Build reusable model libraries that become the canonical source of truth for every future engineering domain Archie will master; write clean, literate Julia code, craft exhaustive test suites, and document assumptions so that the next engineer can extend a heat-exchanger model as easily as importing a package.
• • Surface actionable insights for industrial customers by translating complex thermofluid or electromechanical behaviors into intuitive visual diagnostics; create dashboards that let plant operators see how a fouled cooling-tower fill or an unbalanced three-phase load will cascade through the system hours before failure.
• • Champion a culture of scientific rigor and radical candor—present weekly “sim-to-real” gap analyses, challenge assumptions with data, and celebrate elegant derivations as much as elegant code; your credibility will shape both Archie’s capabilities and the company’s technical brand.
• • Contribute to open-source MTK and SciML ecosystems, upstreaming performance patches and new component models so that the broader Julia community benefits from the breakthroughs you create; speak at conferences, publish white-papers, and turn P-1 AI into the de-facto leader in differentiable engineering simulation.
• • Embrace the velocity of an early-stage startup: context-switch between debugging a stiff ODE at 9 a.m., debating equity allocation at noon, and white-boarding a new product feature at 5 p.m.; wear multiple hats without ego, ship code that same night, and wake up the next morning ready to do it again.
• • Spend one week out of every six co-located in our San Francisco headquarters (or relocate permanently if desired), soaking in high-bandwidth hallway conversations, rapid-fire pair programming, and the occasional sunset cruise on the bay—all while knowing that your work is literally teaching an AI to redesign the built world.