*Telecommuting permitted: work may be performed within normal commuting distance from the Red Hat, LLC office in Boston, MA. Lead the design and implementation of high-performance Python, C++, CUDA, and Triton code for vLLM-based inference systems, GPU kernels, and numerical computing components in production-scale LLM serving environments. What You Will Do: • Architect, develop, and optimize advanced LLM inference algorithms, including quantization, sparsification, speculative decoding, paged attention, KV cache optimization, batching, scheduling, and memory management, to significantly improve latency, throughput, scalability, and cost efficiency. • Drive deep performance profiling, benchmarking, and modeling on NVIDIA GPUs, using tools such as Nsight Systems, Nsight Compute, CUPTI, PyTorch Profiler, CUDA, Triton, and CUTLASS to identify bottlenecks and deliver measurable system-level improvements. • Own the end-to-end optimization of vLLM-based inference pipelines, including kernel-level tuning, runtime optimization, distributed inference, multi-GPU serving, tensor/pipeline parallelism, and large-scale deployment performance. • Evaluate, prototype, and integrate state-of-the-art LLM serving techniques from cutting-edge research and open-source communities, including quantization, sparsity, speculative decoding, continuous batching, prefix caching, attention optimization, and model compression. • Set engineering standards for performance-critical code, conduct rigorous peer reviews, guide testing and benchmarking strategies, and leverage AI-assisted development tools to maintain high code quality, reliability, and maintainability. • Contribute to and maintain upstream open-source ML infrastructure vLLM, including reviewing pull requests, triaging issues, participating in design discussions, and supporting release readiness. • Stay at the forefront of LLM systems research and GPU computing advancements, proactively identifying opportunities to improve the product through emerging techniques in serving efficiency, hardware utilization, and model acceleration. What You Will Bring: • Master's degree (U.S. or foreign equivalent) in Computer Science, Computer Engineering, or related field and three (3) years of experience in the job offered or related role OR Bachelor's degree (U.S. or foreign equivalent) in Computer Science, Computer Engineering, or related field and five (5) years of experience in the job offered or related role. • Must have three (3) years of experience with: high-performance Python and Modern C++ development; performance analysis, benchmarking, and profiling on NVIDIA GPUs using Nsight or similar tools; and GPU memory architecture and memory optimization for deep learning workloads. • Must have two (2) years of experience with: making open-source pull requests to PyTorch or vLLM; CUDA, Triton, or CUTLASS kernel optimization; LLM inference optimization, including quantization, KV cache or batching; distributed computing including inter and intra GPU communication protocols; and leading technical design, reviewing performance-critical code. #LI-DNI The salary range for this position is $186,160 - $287,410/year. Actual offer will be based on your qualifications. Pay Transparency Red Hat determines compensation based on several factors including but not limited to job location, experience, applicable skills and training, external market value, and internal pay equity. Annual salary is one component of Red Hat’s compensation package. This position may also be eligible for bonus, commission, and/or equity. For positions with Remote-US locations, the actual salary range for the position may differ based on location but will be commensurate with job duties and relevant work experience. About Red Hat Red Hat is the world’s leading provider of enterprise open source software solutions, using a community-powered approach to deliver high-performing Linux, cloud, container, and Kubernetes technologies. Spread across 40+ countries, our associates work flexibly across work environments, from in-office, to office-flex, to fully remote, depending on the requirements of their role. Red Hatters are encouraged to bring their best ideas, no matter their title or tenure. We're a leader in open source because of our open and inclusive environment. We hire creative, passionate people ready to contribute their ideas, help solve complex problems, and make an impact. Inclusion at Red Hat Red Hat’s culture is built on the open source principles of transparency, collaboration, and inclusion, where the best ideas can come from anywhere and anyone. When this is realized, it empowers people from different backgrounds, perspectives, and experiences to come together to share ideas, challenge the status quo, and drive innovation. Our aspiration is that everyone experiences this culture with equal opportunity and access, and that all voices are not only heard but also celebrated. We hope you will join our celebration, and we welcome and encourage applicants from all the beautiful dimensions that compose our global village. Equal Opportunity Policy (EEO) Red Hat is proud to be an equal opportunity workplace and an affirmative action employer. We review applications for employment without regard to their race, color, religion, sex, sexual orientation, gender identity, national origin, ancestry, citizenship, age, veteran status, genetic information, physical or mental disability, medical condition, marital status, or any other basis prohibited by law. Red Hat does not seek or accept unsolicited resumes or CVs from recruitment agencies. We are not responsible for, and will not pay, any fees, commissions, or any other payment related to unsolicited resumes or CVs except as required in a written contract between Red Hat and the recruitment agency or party requesting payment of a fee. Red Hat supports individuals with disabilities and provides reasonable accommodations to job applicants. If you need assistance completing our online job application, email application-assistance@redhat.com. General inquiries, such as those regarding the status of a job application, will not receive a reply.