Tech Stack
Tag name is followed by "@" symbol and proficiency level value.
About proficiency levels:
- 1-2 — basic awareness. Minimal hands-on experience, and a rudimentary understanding of the technology's purpose;
- 3-6 — daily use. Comfortable and regular usage, capable of handling common tasks and challenges related to the technology;
- 7-9 — you are an expert, you can teach others, you know all the pitfalls and tricks;
- 10 — exceptional knowledge, comprehensive understanding, and adeptness in all aspects of the technology, including advanced problem-solving. Think twice before claiming or demanding such level.
AI
Deep Learning @ 5
Machine Learning @ 3
PyTorch @ 5
Python @ 5
- 1-2 — basic awareness. Minimal hands-on experience, and a rudimentary understanding of the technology's purpose;
- 3-6 — daily use. Comfortable and regular usage, capable of handling common tasks and challenges related to the technology;
- 7-9 — you are an expert, you can teach others, you know all the pitfalls and tricks;
- 10 — exceptional knowledge, comprehensive understanding, and adeptness in all aspects of the technology, including advanced problem-solving. Think twice before claiming or demanding such level.
Details
NVIDIA's Relational Foundation Model team is building a unified foundation model that understands the structure, context, and relationships within relational databases and heterogeneous graphs. The work involves designing, building, and evaluating novel Transformer and graph neural network architectures that generalize across diverse data schemas. Applications include recommendation systems, demand forecasting, fraud detection, predictive maintenance, forecasting, entity matching, and customer retention.
The role covers the full machine learning lifecycle, including architecture exploration, large-scale training, post-training optimization, and high-performance inference. NVIDIA foundation models are optimized for NVIDIA-accelerated infrastructure and are intended to translate advanced AI research into deployable production systems.
Responsibilities
- Collaborate with researchers and engineers to enhance Transformer- and GNN-based models for relational schemas and heterogeneous graphs.
- Contribute to high-impact use cases including forecasting, entity matching, customer retention, and fraud detection.
- Apply machine learning and artificial intelligence expertise to develop scalable and adaptable solutions.
- Work across the full lifecycle of modern machine learning systems, from architecture design and training to post-training optimization and inference acceleration.
- Contribute to the next generation of the Relational Foundation Model.
Requirements
- Master's or PhD degree in Machine Learning, Computer Science, or equivalent experience.
- Proficiency in Python and deep learning frameworks such as PyTorch.
- At least 8 years of research experience designing machine learning algorithm solutions.
- Practical experience using predictive models in real-world applications.
Preferred Qualifications
- Familiarity with graph-based machine learning.
- Publications at venues such as NeurIPS, ICLR, ICML, or similar.
Benefits
- Equity and benefits are provided.
- NVIDIA is committed to fostering an inclusive work environment and is an equal opportunity employer.
Applications will be accepted at least until August 21, 2026. This posting is for an existing vacancy.