About the job
About Us:
Axiomatic AI is pioneering a new generation of AI systems that embody the principles of the scientific method. By merging deep learning with formal logic and physics-based modeling, we are creating verifiable and interpretable AI systems that enhance and support human researchers in critical scientific and engineering tasks.
Our ambitious mission, 30×30, aims to achieve a 30-fold improvement in the speed, accessibility, and cost-effectiveness of semiconductor and photonic hardware development by the year 2030.
We are set to transform hardware design and simulation across these sectors, assembling a team of passionate professionals dedicated to translating innovative research into commercial products.
Position Overview:
As a Research Software Engineer with a focus on scientific computing systems, you will be responsible for constructing and scaling the computational foundation of our scientific tools. This includes accelerating simulation and optimization workloads, facilitating robust distributed execution, and ensuring correctness and reproducibility across numerical pipelines. You will collaborate closely with a diverse team comprising AI Engineers, Software Engineers, Mathematicians, and Physicists to develop tools that revolutionize how engineers and scientists leverage AI in their daily operations.
Your Mission:
- Scalable Scientific Computing: Develop high-performance, reliable systems for simulation, inference, optimization, and uncertainty quantification workflows, particularly in the realms of electromagnetic simulation and inverse design.
- Performance Optimization: Analyze and enhance performance throughout end-to-end pipelines (CPU/GPU).
- Distributed Execution & HPC: Design and sustain distributed computing infrastructure for extensive sweeps and multiple experimental runs (multi-GPU/multi-node), emphasizing reproducibility, observability, and developer usability.
- Verification & QA: Establish rigorous testing and verification protocols for scientific computing pipelines (numerical regression, invariants, convergence tests, golden datasets), ensuring reliable results over time.

