Used Tools & Technologies
Not specified
Required Skills & Competences ?
Python @ 6 Machine Learning @ 4 Communication @ 7 LLM @ 4 PyTorch @ 4 GPU @ 4Details
NVIDIA is seeking a Senior Machine Learning software engineer to discover and innovate new low-precision and sparsity recipes in the pretraining setting. You will develop next-generation software to use novel hardware features on current GPUs and provide guidance for next-gen GPU feature design. The role spans recipe design for all phases of the LLM life cycle: pre-training, post-training, and generation. Work will contribute to software productization in libraries such as Megatron-LM, Transformer Engine, cuDNN, and cuBLAS.
Responsibilities
- Keep abreast of quantized LLM training research.
- Build robust and reproducible training recipes.
- Collaborate closely with hardware, software, and research teams to assess and adopt deep learning algorithmic advancements in quantization.
- Work with production software teams to realize recipes in production workflows.
Requirements
- PhD or M.S. degree (or equivalent experience) in Computer Science or a related field, and 5+ years of relevant software engineering experience.
- Proficient in Python.
- Experience with PyTorch or a similar framework.
- Solid foundation in LLM pre-training, post-training, or generation.
- Proficient in the math of machine learning.
- Strong written and oral communication skills.
Ways to stand out / Nice to have
- Proficient in precision and numerics for ML.
- Familiarity with FP8 and MX formats for training.
- Strong programming skills and ability to debug ML systems.
Benefits and other information
- Base salary range (varies by level and location): USD 184,000 - 287,500 for Level 4; USD 224,000 - 356,500 for Level 5.
- Eligible for equity and company benefits.
- Applications accepted at least until September 21, 2025.
- NVIDIA is an equal opportunity employer and committed to fostering a diverse work environment.