AI/ML Engineer – Infrastructure & Performance Optimization

Seeking an AI/ML Engineer to join a team focused on developing high-performance software and hardware technologies that power cutting-edge AI platforms. This role centers on building tools and systems that extract, profile, and analyze AI workloads to optimize performance across custom and next-generation AI hardware. Key Responsibilities: • Extract AI/ML operators (e.g., ATen, Triton) from models for detailed analysis. • Execute operators on various devices and gather comprehensive performance metrics. • Process and store collected data in structured databases while maintaining data quality and integrity. • Develop and maintain APIs and web interfaces to enable efficient querying and visualization of performance data. • Collaborate with cross-functional teams to define technical scope, requirements, and project goals. • Support integration of performance data pipelines into automated CI workflows. • Participate in the evaluation and optimization of AI operator execution on proprietary hardware. Required Skills: • Proficient in production-level Python programming. • Deep understanding of PyTorch internals, including dispatcher and Kineto trace. • Experience working with CUDA and Triton kernel development. • Skilled in SQL and database management systems. • Comfortable working in Linux environments and scripting with Bash. • Self-motivated and capable of working independently on complex systems. Preferred Skills: • Experience with Large Language Models (LLMs), particularly LLaMA. • Knowledge of CI/CD systems and test automation practices. • Background in profiling and optimizing machine learning workloads.