SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 250,000 - 350,000 / annual
Periodic Labs is building AI systems that simulate physical science, verify predictions, and train on the full scientific method. As a Computational Scientist, you will develop differentiable, accelerator-ready simulations for industrially relevant continuum-physics problems, combining deep expertise in governing equations, solver implementation, and deep learning.
You will build and extend differentiable solvers for continuum simulation, including fluid dynamics and multi-scale, multi-physics problems. Core responsibilities include implementing numerical methods from first principles, diagnosing convergence and stability issues, and combining simulation with deep learning for surrogate modeling, inverse problems, and optimization. You'll leverage automatic differentiation and modern accelerators (JAX, PyTorch) to make simulations scalable and trainable, validate models against experiments and benchmarks, and create datasets to guide LLM development for automating these tasks.
Ideal candidates hold a PhD or equivalent research experience in applied mathematics, computational science, physics, engineering, or computer science. You should have code-level experience building or substantially modifying PDE solvers and differentiable simulations, deep expertise in at least one continuum domain (especially fluid dynamics), and meaningful experience training deep-learning models for physical systems. Strong Python and software-engineering skills are essential, particularly with JAX, PyTorch, Julia, or C++. You should have applied simulation to realistic scientific or engineering problems beyond academic benchmarks and bring a startup mentality with ownership and good judgment under uncertainty.
Strong additional qualifications include expertise in multiple continuum domains or multiphysics modeling, knowledge of adjoint methods and differentiable programming, GPU/TPU acceleration experience, contributions to scientific open-source software, and experience connecting simulation to experiments or engineering workflows.
