Manager, System Software Engineering - Local AI

NVIDIA · India, Pune

onsitefull-time6-10 years

posted 25 Jul

NVIDIA has continuously reinvented itself for over two decades. The invention of the GPU in 1999 propelled the growth of PC gaming, redefined modern computer graphics, and revolutionized parallel computing. More recently, GPU deep learning has ignited modern AI, positioning NVIDIA as a leading AI computing company. There is a growing focus on delivering AI models locally, closer to the source of data. This reduces latency, improves real-time processing, and addresses privacy concerns by minimizing data transfer to centralized servers. As technology advances, client-side AI (local execution) will play a key role in crafting digital experiences. The Local AI team is seeking a System Software Man a g e r to lead development of an efficient on-device AI software stack. The software stack will support RTX, RTX Pro, and DGX-class systems. This role focuses on high-perfor man ce local inference, agentic workloads, low latency, efficient memory use, scalable infrastructure, practical deployment on resource-constrained platforms, and delivering a streamlined out-of-box experience for developers and end users. What you'll be doing: Lead and grow a team building the on-device AI inference platform for RTX, RTX Pro, and DGX GPUs, with accountability for execution, technical direction, delivery quality, and roadmap alignment. Drive cross-functional alignment with NVIDIA’s software, research, architecture, and product teams, along with industry partners and open-source communities, to build strategy and strengthen the AI ecosystem across RTX and DGX platforms. Provide technical leadership for the architecture and evolution of modern inference runtimes and execution stacks across frameworks such as Llama.cpp, vLLM , PyTorch , WinML , DXCGC, and TensorRT-RTX, spanning workloads including LLMs, vision-language models, TTS, ASR, and diffusion models. Mentor engineers, develop technical leaders, and foster a high-perfor man ce team culture centred on innovation, collaboration, and operational excellence. Coordinate end-to-end optimization of AI models, data pipelines, and inference runtimes to improve perfor man ce across current and next-generation GPU architectures. Drive adoption of model optimization techniques such as quantization, pruning, sparsity, and distillation to enable efficient deployment of large models on local and edge devices. Establish team processes for system-level debugging, perfor man ce optimization, and perfor man ce-accuracy trade-off analysis, including infrastructure for perfor man ce and accuracy sweeps, gap analysis, and production-readiness improvements. What we need to see: 5+ overall years of industry experience and 2+ years of engineering leadership experience, combined with a Bachelor’s, Master’s, or PhD in Computer Science, Software Engineering, Mathematics, or a related field. Proven experience leading high-performing engineering teams in systems software, AI infrastructure, inference runtimes, or related domains. Strong technical foundation in C++ software development, debugging, data structures, algorithms, and machine learning systems. Extensive background in AI inference pipelines and Deep Learning frameworks like Llama.cpp, vLLM , PyTorch , WinML , DXCGC, and TensorRT. Deep understanding of inference backends and runtime internals, including scheduling, memory man a g e ment, KV-cache behavior , graph execution, quantization, and hardware-aware optimization. Strong analytical and problem-solving skills, with the ability to balance technical depth, execution speed, and organizational priorities in a fast-paced environment. Excellent written and verbal communication skills, with proven ability to collaborate across engineering, product, research, and executive collaborators. Ways to stand out from the crowd : Strong understanding of modern machine learning, deep neural networks, and generative AI, along with contributions to notable open-source projects. Demonstrated success building teams,