Principal Systems Software Engineer, Semiconductor Systems Inspection

What you'll do

• Define and prototype AI system architectures for optical and e-beam inspection, metrology, and process data.
• Advance world foundation model capabilities for vision, time-series, and multimodal representation learning in semiconductor manufacturing.
• Develop workflows for defect detection, classification, localization, segmentation, and equipment time-series anomaly detection.
• Establish evaluation methods to measure model quality and robustness under tool-to-tool and fab-to-fab domain shifts.
• Design fab-local evaluation and adaptation workflows that preserve customer data sovereignty.
• Connect inspection results with equipment signals and process history to support manufacturing-wide root-cause analysis.
• Design agentic workflows for air-gapped fabs to connect data triage, model inference, and human approval.
• Optimize and package inference capabilities to meet production latency, memory, throughput, and security budgets.

What we're looking for

• MS or PhD in Computer Science, Electrical Engineering, Computer Engineering, or a related technical field, or equivalent experience.
• 12+ years of professional software or algorithm engineering experience, including technical ownership and architectural direction.
• 8+ years of experience in semiconductor manufacturing, including inspection, metrology, defect review, process control, yield engineering, or fab-data systems.
• 4+ years of hands-on deep learning, machine learning, computer vision, or applied AI experience within the last 24 months.
• Hands-on delivery of software that shipped in semiconductor equipment, operated in a fab, or was adopted by a manufacturing customer.
• Strong Python skills and production experience with PyTorch or TensorFlow.
• Working GPU literacy to reason about inference performance and evaluate hardware and software bottlenecks.
• Experience adapting large pretrained vision, multimodal, or world foundation models to specialized industrial domains (preferred).