Senior Performance and Scale Engineer - Distributed LLM Inference

This role needs a seasoned engineer that thinks creatively, adapts to rapid change, and has the willingness to learn and apply new technologies. You will be joining a vibrant open source culture, and helping promote performance and innovation in this company engineering team. The border mission of the Performance and Scale team is to establish performance and scale leadership of the company product and cloud services portfolio. The scope includes component level, system and solution analysis and targeted enhancements. The team collaborates with engineering, product management, product marketing and customer support as well as our companys hardware and software ecosystem partners.
At our company, our commitment to open source innovation extends beyond our products - its embedded in how we work and grow. workers embrace change especially in our fast-moving technological landscape and have a strong growth mindset. That's why we encourage our teams to proactively, thoughtfully, and ethically use AI to simplify their workflows, cut complexity, and boost efficiency. This empowers our associates to focus on higher-impact work, creating smart, more innovative solutions that solve our customers' most pressing challenges.
What you will do:
Define and track key performance indicators (KPIs) and service level objectives (SLOs) for large-scale, distributed LLM inference services in Kubernetes/OpenShift
Participate in the performance roadmap for distributed inference, including multi-node and multi-GPU scaling studies, interconnect performance analysis, and competitive benchmarking
Formulate performance test plans and execute performance benchmarks to characterize performance, drive improvements, and detect performance issues through data analysis and visualization
Develop and maintain tools, scripts, and automated solutions that streamline performance benchmarking tasks.
Collaborate with cross-functional engineering teams to identify and address performance issues.
Partner with DevOps to bake performance gates into GitHub Actions/OpenShift Pipelines.
Explore and experiment with emerging AI technologies relevant to software development, proactively identifying opportunities to incorporate new AI capabilities into existing workflows and tooling.
Triage field and customer escalations related to performance; distill findings into upstream issues and product backlog items.
Publish results, recommendations, and best practices through internal reports, presentations, external blogs, and official documentation.
Represent the team at internal and external conferences, presenting key findings and strategies.