Senior Foundation Model Systems Engineer

Jobtailor

California (MO)

On-site

USD 180,000 - 260,000

Full time

14 days+

Get more replies from employers

Send a job-specific resume in minutes.

Job summary

NVIDIA is seeking a highly skilled researcher/engineer to advance foundation models, train and deploy multimodal data pipelines, and optimize large-scale AI systems. The role emphasizes building scalable infrastructure, reducing costs, and enabling efficient inference across cloud and edge environments, with extensive work on Megatron-LM, Megatron Bridge, and NeMo-RL.

Ideal candidates have 5+ years of experience, strong ML/DL foundations, and hands-on PyTorch development in distributed training

Qualifications

MS or PhD in Computer Science or related field (or equivalent experience).
5+ years of relevant industry experience.
Strong foundation in machine learning, deep learning, and optimization.
Excellent software engineering skills, including Python and PyTorch.
Experience building high-performance software for large-scale AI systems.
Experience with distributed training at scale, such as Megatron-LM, Megatron Bridge, FSDP, and TP/PP/CP/DP.
Experience with optimizer research and efficient sparse or long-context attention.
Experience with supervised fine-tuning, reinforcement learning for LLMs, PPO, GRPO, asynchronous RL, or NeMo-RL.
Experience with model compression, quantization, pruning, knowledge distillation, neural architecture search, or diffusion/non-autoregressive language models.
Experience contributing to open-source AI frameworks such as Megatron-LM, Megatron Bridge, NeMo-RL, or Hugging Face Transformers
Experience with GPU performance optimization, distributed systems, latency/throughput analysis, and profiling of large-scale AI workloads

Responsibilities

Advance the state of the art in foundation model development, training, and deployment.
Design scalable systems for preparing high-quality multimodal datasets for frontier foundation model training.
Develop algorithms and systems that improve scalability, efficiency, and cost of large-scale pre-training and post-training.
Advance techniques that improve inference performance, reduce deployment cost, and enable efficient serving across cloud and edge platforms.
Develop reusable infrastructure and contribute new model support to NVIDIA's open-source GenAI training platform.
Collaborate with research, product, and infrastructure teams to design new algorithms and optimize existing systems.
Contribute to NVIDIA's open-source AI stack, including Megatron-LM, Megatron Bridge, and NeMo-RL