Unconventional AI
System Modeling (Dynamic Systems Simulation)
Mountain View, CA I US Remote +2 · Remote · Full-time · Posted Apr 1, 2026
- Location
- Mountain View, CA I US Remote
Mountain View
Remote - Experience
- Proven industry experience in high-fidelity circuit simulation; no number of years stated. Also requests strong ML implementation and systems optimization experience.
- Education
- MS/PhD in Electrical Engineering, Computer Engineering, or a closely related field, or BS with substantial evidence of equivalent research/engineering depth in circuit simulation.
About the job
Develop high-speed, high-fidelity simulation frameworks and PyTorch or JAX components for physics-based, time-varying circuit systems used in AI computing.
What you'll do
- Build custom behavioral simulators that preserve key dynamics while enabling faster iteration than traditional circuit simulators.
- Develop models spanning neural networks, circuit behavior, and governing differential equations.
- Contribute to GPU-accelerated ML simulations and cross-layer system integration.
What they ask for
- Knowledge of dynamical systems simulation and analog/mixed-signal circuit design, including transistor-level principles and nonidealities.
- Proficiency in PyTorch or JAX, including custom autograd functions and numerical solvers for dynamic processes.
- Strong understanding of linear and nonlinear dynamics, state-space representations, and differential equations in an ML context.
- Proven industry experience building high-fidelity circuit simulations that balance computational efficiency and physical accuracy.
Nice to have
- Strong Python engineering skills, including modular design, testing, packaging, and CI.
- Experience with GPU programming (such as CUDA or Triton), PyTorch internals, or distributed/HPC patterns.
Tools
PyTorch, JAX, CUDA, Triton, NumPy, TensorFlow, Modal, SciPy
AI focus
Develop simulations and models for physics-based computing systems and next-generation AI architectures, including GPU-accelerated ML system simulations.
Model-written summary · the original posting is authoritative