Modelling Engineer (Mid-Staff)

Build the Future of AI with Optical Compute
We're pioneering optical processors to train and run inference on large-scale AI models. Join a team of highly motivated and skilled engineers dedicated to rapid innovation and high-impact outcomes.

We're looking for a Senior Performance Modelling Engineer to design and own the analytical and simulation models that guide the evolution of our Optical TPU (OTPU) architecture and software. You will be instrumental in building functional and high-fidelity, cycle-accurate models of our optical compute system.

This is a high-leverage role that sits at the intersection of hardware design, software tooling, and machine learning workloads-ideal for engineers who thrive on data-driven decisions, rapid iteration, and solving complex performance challenges.

• End-to-End Ownership: Lead and deliver critical projects that enable major technical and business milestones.

• Cross-Functional Collaboration: Work closely with hardware, compiler, and ML framework teams to ensure performance models are both accurate and actionable.

• Simulator Development:

  • Build and maintain a functional simulator of the OTPU pipeline and subsystems.

  • Develop architectural and cycle-accurate simulators to identify bottlenecks and optimize throughput, latency, and utilization.

• Benchmarking & Bottleneck Analysis: Instrument LLMs, diffusion models, and graph workloads to generate detailed traces for deep performance analysis.

• Design-Space Exploration: Run extensive parameter sweeps to explore architectural tradeoffs, and deliver clear, quantitative insights that guide our hardware, software, and optical designs.

• Tooling & Automation: Create robust Python/C++ tools for trace parsing, statistical analysis, and visualization. Integrate models into CI pipelines for automated performance regression testing.

• 5+ years of experience building performance or power models for CPUs, GPUs, ASICs, or custom accelerators.

• Proficiency in C++ and Python, with hands-on experience in developing discrete-event or cycle-accurate simulators (eg, gem5, SystemC, or custom tools).

• Strong understanding of computer architecture fundamentals: memory systems, interconnects, queuing theory, Amdahl's and Gustafson's laws.

• Familiarity with machine learning workloads and frameworks like PyTorch, TensorFlow, or JAX.

• Ability to interpret RTL/schematics and discuss micro-architectural trade-offs with hardware engineers.

• Excellent data visualization and communication skills - capable of distilling millions of simulation samples into a single, decisive insight.

• Advanced degree (Master's or PhD) in Electrical Engineering, Computer Science, Physics, Applied Math, or a related field.

• Open-source or personal projects involving simulators, ML kernels, or performance analysis.

This role offers a unique opportunity to shape the direction of optical compute at a foundational level. If you're excited to work at the cutting edge of hardware, software, and AI, we'd love to hear from you.