About the Role
NVIDIA is seeking a Senior Deep Learning Algorithm Engineer for their core AI Frameworks (Megatron Core and NeMo Framework) team. This role involves designing, developing, and optimizing diverse real-world workloads. Megatron Core and NeMo Framework are open-source, scalable, and cloud-native frameworks designed for researchers and developers working on Large Language Models (LLM) and Multimodal (MM) foundation model pretraining and post-training. These GenAI Frameworks offer end-to-end model training, including pretraining, reasoning, alignment, customization, evaluation, deployment, and tooling for performance optimization and enhanced user experience.
In this critical position, you will expand the capabilities of Megatron Core and NeMo Framework, empowering users to develop, train, and optimize models. This includes designing and implementing the latest distributed training algorithms, model parallel paradigms, and model optimizations, defining robust APIs, meticulously analyzing and tuning performance, and expanding toolkits and libraries to be more comprehensive and coherent. You will collaborate with internal partners, users, and the open-source community to analyze, design, and implement highly optimized solutions.
What you’ll be doing:
- Develop algorithms for AI/DL, data analytics, machine learning, or scientific computing.
- Contribute to and advance open-source projects like NeMo-RL, Megatron Core, and NeMo Framework.
- Solve large-scale, end-to-end AI training and inference challenges across the entire model lifecycle, from initial orchestration and data pre-processing to model training, tuning, and deployment.
- Work at the intersection of computer architecture, libraries, frameworks, AI applications, and the entire software stack.
- Innovate and improve model architectures, distributed training algorithms, and model parallel paradigms.
- Conduct performance tuning and optimizations, and perform model training and finetuning with mixed precision recipes on next-gen NVIDIA GPU architectures.
- Research, prototype, and develop robust and scalable AI tools and pipelines.
What we need to see:
- MS, PhD or equivalent experience in Computer Science, AI, Applied Math, or related fields.
- 5+ years of industry experience.
- Experience with AI Frameworks (e.g., PyTorch, JAX, Ray), and/or inference and deployment environments (e.g., TRTLLM, vLLM, SGLang).
- Proficient in Python programming, software design, debugging, performance analysis, test design, and documentation.
- Consistent record of working effectively across multiple engineering initiatives and improving AI libraries with new innovations.
- Strong understanding of AI/Deep Learning fundamentals and their practical applications.
Ways to stand out from the crowd:
- Hands-on experience in large-scale AI training, with a deep understanding of core compute system concepts (such as latency/throughput bottlenecks, pipelining, and multiprocessing) and demonstrated excellence in related performance analysis and tuning.
- Prior experience with Reinforcement Learning algorithms and compute patterns.
- Expertise in distributed computing, model parallelism, and mixed precision training.
- Prior experience with Generative AI techniques applied to LLM and Multi-Modal learning (Text, Image, and Video).
- Knowledge of GPU/CPU architecture and related numerical software.