Job Specifications
About Us
TAngraFX is at the forefront of blending cutting-edge computational physics with artificial intelligence to solve the world's most complex engineering and scientific challenges. From designing next-generation materials and energy systems to pushing the boundaries of climate science and aerospace, we use massive computing power and innovative algorithms to simulate reality. Join us to build the future of simulation.
The Role: HPC AI Developer
We are seeking a passionate and skilled HPC AI Developer to join our elite research and engineering team. You will be at the heart of our mission to integrate state-of-the-art AI and Machine Learning models with traditional High-Performance Computing (HPC) simulations.
Your primary focus will be on developing, scaling, and optimizing AI-driven surrogate models (like Graph Neural Networks, Fourier Neural Operators, or Physics-Informed Neural Networks) to accelerate or even replace computationally intensive physics simulations. You will work directly with domain scientists and engineers to turn groundbreaking research into production-ready, high-performance code that runs on the world's largest supercomputers and GPU clusters.
Key Responsibilities
Research & Development: Design, implement, and train novel AI/ML models for physics simulation, with a focus on learning the underlying dynamics of physical systems.
HPC Integration & Scaling: Deploy, scale, and optimize these AI models on large-scale HPC and GPU clusters, leveraging frameworks like PyTorch (with Distributed Data Parallel - DDP), JAX, or TensorFlow.
Performance Optimization: Profiling and extreme performance tuning of AI workloads, from single-GPU kernel optimization using CUDA/Triton to multi-node, multi-GPU training and inference.
Bridge Physics & AI: Collaborate with domain experts to understand the fundamental physics and numerical methods, ensuring our models are both computationally efficient and physically consistent.
Software Engineering: Develop robust, maintainable, and scalable scientific software in Python and C++. Contribute to our core codebase and MLOps pipelines.
Stay Cutting-Edge: Keep abreast of the latest developments in scientific machine learning (SciML), HPC, and AI hardware to continuously improve our capabilities.
What We're Looking For (Qualifications)
Required:
M.S. or Ph.D. in Computer Science, Computational Physics, Aerospace Engineering, Applied Mathematics, or a related field, with a strong focus on AI/ML and HPC.
3+ years of proven experience developing and training deep learning models, with a strong portfolio or publications in relevant areas (e.g., GNNs, PINNs, Operator Learning).
Strong proficiency in Python and deep learning frameworks (PyTorch strongly preferred).
Hands-on experience with parallel programming and scaling ML models on HPC systems or large-scale cloud GPU environments (e.g., using SLURM, MPI, NCCL).
Solid understanding of software engineering best practices (version control, CI/CD, testing).
A passion for applying AI to solve real-world scientific problems.
Highly Desirable (Nice-to-Have):
Experience with performance profiling and optimization tools (Nsight Systems/Compute, VTune).
Knowledge of low-level GPU programming (CUDA, OpenCL, or Triton).
Background in traditional computational physics methods (FEA, CFD, Molecular Dynamics).
Familiarity with scientific data formats (HDF5, NetCDF) and visualization.
Experience with MLOps tools (Kubernetes, Docker, MLflow, Weights & Biases).
About the Company
At TAngraFX, we're revolutionizing the world of VFX by providing cutting-edge multi-physics engines that integrate seamlessly with Blender, with more software compatibility coming soon. Our Software: TAngra - The Multi-Physics Engine for Blender, Maya, 3ds Max: Are you tired of the limitations of traditional VFX tools? Want a solution that supports a range of advanced simulations to bring your vision to life? With TAngra, you get: MPM (Material Point Method) for realistic simulations of granular materials like sand and snow....
Know more