Job Description

NVIDIA seeks a motivated AI Researcher to join the BioNemo/TAO R&D team in Hanoi or Ho Chi Minh city. In this position, you will invent, develop, and test innovative deep learning and generative AI models — such as diffusion models, graph neural networks, and large-scale foundation models — targeting critical scientific problems in drug discovery, urban planning, and warehouse operations. You will work where advanced AI research meets life sciences and smart environments, building models that extend AI's potential to improve human health and intelligent urban warehouses.

Collaborating with top researchers, engineers, and diverse teams, you will develop NVIDIA's BioNemo/TAO platform further. You will transform innovative research into production-ready solutions coordinated across NVIDIA SDKs, Inference Microservices (NIM), and practical scientific workflows. This position offers scientists a chance to use AI to accelerate science and positively affect human health.

What You'll Be Doing

Investigate, design, and optimize deep learning models — including diffusion models, flow-matching networks, transformer architectures, and graph neural networks — for structured prediction tasks in molecular and protein science.
Build and implement generative AI systems for various user cases. These include molecular building, structural prediction, and image/video abnormal generation. The focus is on producing reliable, ranked outputs for scientific and real-world use.
Contribute to the development and scaling of foundation models for several domain applications, including protein structure, molecular interaction, and other fields.
Apply and adapt NVIDIA acceleration techniques (e.g., CuEquivariance, efficient attention, mixed-precision inference) to deploy large-scale AI models on NVIDIA GPU hardware with high efficiency and accuracy.
Conduct detailed experiments, ablation studies, and benchmarking to evaluate model accuracy, robustness, generalization, and scalability across diverse datasets.
Work together with multi-functional groups to incorporate research findings into the BioNemo SDK, NVIDIA production pipelines (e.g., NVIDIA Inference Microservices — NIM), and customer-facing scientific workflows.
Build and improve post-processing and evaluation pipelines that transform raw model predictions into actionable insights for downstream tasks.
Participate actively in research discussions, paper reading groups, code reviews, and technical documentation to share knowledge and elevate team methodology.