Principal Software Architect – High Performance Computing

About the role

As a Principal Software Architect in High Performance Computing, you will lead the design and implementation of scalable infrastructure that combines CPUs, GPUs and FPGAs for semiconductor equipment. You will work on low‑latency, high‑throughput image processing and deep‑learning workloads that drive chip manufacturing process control.

Key responsibilities

• Design and implement robust, scalable infrastructure solutions across heterogeneous processors.
• Analyze and partition workloads to appropriate compute units, optimizing AI inference and parallel processing.
• Collaborate with algorithm engineers, product managers and business stakeholders to translate requirements into architectural designs.
• Develop prototypes and code to validate designs and unblock engineering teams.
• Profile clusters and nodes, conduct performance tuning, capacity planning and monitor GPU metrics.

Required profile

• 12–18 years of experience building scalable, secure infrastructure with CPUs, GPUs and FPGAs.
• Deep expertise in parallel programming (CUDA) and AI inference servers such as NVIDIA Triton.
• Strong knowledge of C/C++, data structures, algorithms and complexity analysis.
• Experience with distributed HPC software (MPI, UCX) and GPU programming frameworks.

Required skills

• C/C++
• CUDA
• OpenMP
• OpenACC
• OpenCL
• MPI
• UCX
• NVIDIA Triton
• VTune
• Nsight
• perf
• Docker
• Singularity

What we offer

• Supportive work culture with learning and development opportunities.
• Benefits focused on health, wellbeing and work‑life balance.
• Opportunities to work on cutting‑edge AI and machine learning technologies.