Senior Software Engineer, Inference

Senior Software Engineer, Inference    This role has been designed as 'Hybrid' with a requirement that you will work on average 2 days per week from an HPE office.

Hewlett Packard Enterprise is the global edge-to-cloud company advancing the way people live and work. We help companies connect, protect, analyze, and act on their data and applications wherever they live, from edge to cloud, so they can turn insights into outcomes at the speed required to thrive in today’s complex world. Our culture thrives on finding new and better ways to accelerate what’s next. We know varied backgrounds are valued and succeed here. We have the flexibility to manage our work and personal needs. We make bold moves, together, and are a force for good. If you are looking to stretch and grow your career our culture will embrace you. Open up opportunities with HPE.

HPE's Private Cloud AI organization is seeking a Senior Software Engineer   to build and evolve the model runtime within HPE AI Essentials, the inference platform used by enterprises to operate large language models on infrastructure they own, including air-gapped and sovereign environments. The core engineering challenge in this domain is not model deployment but sustained execution efficiency: achieving low tail latency and high GPU utilization on customer-owned hardware of varying generation and configuration. In this role you will design and implement key components of that runtime – engine integration, batching, KV cache management, and distributed execution – together with the Kubernetes orchestration layer that supports it. The primary work location is as listed, but could be any other HPE site location in the US ; however, remote work options will be considered. Responsibilities ·        Design, implement, and own major components of the LLM serving deployment, including engine integration, continuous batching, KV cache management and reuse, and quantized execution ·        Partner with inference engineering teams and contribute to improving time-to-first-token, inter-token latency, throughput per GPU, and P95/P99 tail latency ·        Build and operate distributed execution capabilities, including disaggregated prefill/decode, tensor and pipeline parallelism, and KV cache offload across GPU memory, host memory, and RDMA-attached storage ·        Evaluate emerging runtimes, quantization schemes, speculative decoding, and mixture-of-experts serving, and make well-supported recommendations on adoption ·        Contribute to the orchestration layer supporting the runtime, including model admission, GPU scheduling and partitioning, cache-aware request routing, and autoscaling ·        Triage and resolve customer issues end-to-end, identifying root causes and improving systems and processes to prevent recurrence ·        Provide insightful code and design reviews, mentor team members, and lead by example on engineering practices within the team Knowledge and Skills Required ·        Familiar with LLM inference engines such as vLLM, SGLang, TensorRT-LLM, TGI, or NVIDIA NIM, including modification of engine internals ·        Strong understanding of inference internals, including continuous batching, paged attention, KV cache reuse and prefix caching, chunked prefill, quantization, and speculative decoding ·        Working knowledge of tensor and pipeline parallelism, NCCL collective operations, and the GPU memory hierarchy and interconnect characteristics that govern them ·        Advanced proficiency in Kubernetes platform architectures, including operators, custom resources, controllers, and scheduling ·        Strong programming proficiency in Go and Python, with the ability to read, debug, and profile C++/CUDA using tools such as Nsight ·        Familiar with debugging/profiling multi-tier application workloads such as RAG, Agents ·