AI Compiler Engineer

at  Intel

JOB DESCRIPTION

As the AI landscape evolves, the emergence of generative AI has brought the cost of inference into sharp focus, identifying it as a critical bottleneck for innovation and the full utilization of AI capabilities. In the role of Staff AI Compiler Engineer, you’ll be at the forefront of confronting this challenge. By contributing to the development and integration of our state-of-the-art Neural Processing Unit (NPU) compiler within the Windows Machine Learning and DirectML software stack, you will play a central role in offloading cloud inference, thereby revolutionizing the efficiency and scalability of AI applications. With your command of C++ and MLIR, you will not only enhance our NPU and its ecosystem but also redefine the parameters of what’s possible in AI hardware acceleration, paving the way for broader adoption and more sustainable AI innovation. Key responsibilities: - Conduct pioneering research and development in algorithms to push the boundaries of AI efficiency, targeting specific improvements in inference cost and performance. - Implement advanced compilation passes that optimize machine learning models for our NPU, enhancing their performance and reducing latency. - Collaborate closely with hardware design teams to consult on new generation hardware, ensuring that our NPU designs are optimized for the latest AI workloads and inference demands. - Lead research and development efforts for performance optimization, particularly focusing on software-to-hardware mapping for AI models. - Craft and maintain an internal testing framework that sets the bar for industry standards in performance testing. - Track and report engineering progress with precision, promoting transparency and accountability in every aspect of the project lifecycle. - Strategically contribute to the engineering roadmap, aligning with the critical business needs - Maintain a regimen of rigorous code reviews to ensure the highest quality of code. - Embrace incremental development strategies, breaking down complex projects into manageable tasks. - Engage in cross-functional collaboration with peer teams for co-engineering, testing, integration, and complex problem-solving. - Outstanding collaboration and communication abilities. - Experience in managing complex collaborations across multiple teams and stakeholders. - Flexibility to adapt to a rapidly changing tech environment. - Strong commitment to ongoing learning and professional growth.

QUALIFICATIONS

You must possess the below minimum qualifications to be initially considered for this position. Preferred qualifications are in addition to the minimum requirements and are considered a plus factor in identifying top candidates.

MINIMUM QUALIFICATIONS:

• Bachelor’s degree in computer science, electrical engineering, or a STEM related field with 6+ years of experience in three or more the following:
• Solid background in AI frameworks and compiler construction.
• High proficiency in C++ and MLIR, with a track record of applying these skills in a performance-critical environment.
• Agile project management proficiency, with practical experience in tools like GitHub.
• Experience in performance optimization with an emphasis on cost-effective software/hardware mapping.
• Well-versed in testing frameworks and advanced methodologies.

PREFERRED QUALIFICATIONS:

• Master’s degree with 4+ years of experience, OR a PhD with 2+ years of experience in a related discipline.
• Practical experience in Windows driver development.
• In-depth understanding of neural network workloads, with a specialization in AI model optimization for efficient inference.
• Prior involvement with Generative AI integration.

Please refer the Job description for details