AI Frameworks Engineer

Intel's Data Center Group (DCG) develops AI frameworks and workloads that empower researchers and developers to create high-performance, scalable solutions for complex AI challenges, focusing on AI frameworks and workload development for LLM inference on Intel GPUs.

• Bachelor's degree in AI, Computer Science, Computer Engineering, Mathematics, Physics, or related field.
• 4+ years of relevant experience with a Bachelor's degree, or 3+ years with a Master's degree.
• Advanced expertise in Python programming.
• Hands-on experience with AI inference and large language models (LLMs).
• Demonstrated proficiency in PyTorch.
• Solid foundation in AI frameworks, machine learning algorithms, and software engineering practices.
• Strong problem-solving skills with experience in debugging and designing optimized software and algorithms.
• Proficiency in Linux environments from a user perspective.

Nice to have:

• Experience in performance profiling and optimization for AI workloads, including hardware-level profiles.
• Familiarity with parallel and distributed computing concepts and collective communication libraries.
• Familiarity with vLLM or another LLM inference serving framework.
• Track record of accepted contributions to major open-source AI projects.
• Understanding of numerical precision and quantization formats (FP8, INT8, INT4) and accuracy implications.
• Knowledge of writing/optimizing GPU compute kernels.
• Creativity and innovative technical problem-solving skills.
• Strong communication skills and ability to work across global teams.

• Design, develop, and optimize AI framework software for LLM inference on Intel GPUs, contributing upstream to external open-source frameworks such as vLLM and PyTorch.
• Develop and optimize compute kernels for transformer inference.
• Implement and tune distributed execution paths including tensor, pipeline, and data parallelism, collective communication, and asynchronous data movement within deep learning frameworks.
• Transform and optimize the computational-graph representation of neural network models, and develop machine learning primitives in mathematical and kernel libraries.
• Profile inference workloads to identify performance bottlenecks across the stack, quantify them, and propose and drive solutions with component teams.
• Optimize code for various computing hardware backends, adapting to memory hierarchy, instruction sets, and execution models.
• Build and extend unit tests, correctness checks, and performance regression coverage to maintain reliability as upstream projects evolve.
• Collaborate with framework, compiler, and machine learning engineers globally and engage with the open-source community through code review and discussions.

All qualified applicants will receive consideration for employment without regard to legally protected characteristics as described in the Posting Statement.