Nvidia Solution Architect - Agentic AI 

Japan, Tokyo 

593891550

Exploring the latest advancement in model training, fine tuning and customization, while supporting building agentic LLM applications.

Enabling NVIDIA strategic customers to build enterprise AI solutions using accelerated computing stack including NIMs and NeMo microserviecs.

Collaborate with developers and onboard them to NVIDIA AI platforms and services by providing deep technical guidance.

Establishing and building repeatable reference architecture, communicate standard processes and understand solution trade-offs. Share findings and feedback to improve products and services.

Drive pre-sales conversations, build architectures and demos to accelerate the customer AI journey based on NVIDIA products, and work closely with Sales Account Managers to secure design wins.

Create or run Proofs of Concept and demos that require presentation skills, the explanation of complex topics, and Python coding to execute data pipelines, train ML/DL models, and deploy them on container-based orchestrators.

Excellent verbal, written communication, and technical presentation skills in Japanese. Business level English communication is also a requirement.

BS or MS in Computer Science, Engineering, Mathematics, or Physics (or equivalent experience)

5+ years of industry or academic experience related to Generative AI or Deep Learning

Strong coding development and debugging skills. Including experience with Python, C/C++, Bash, and Linux

Ability to multitask effectively in a dynamic environment

Strong analytical and problem-solving skills

Proactive and have a strong desire to share knowledge with clients, partners and co-workers

Expertise in deploying large-scale training and inferencing pipeline

Experience with pre-training, post-training of transformer-based architectures for language or vision

A deep understanding of the latest generative AI or deep learning methods and algorithms

Experience using or operating Kubernetes, as well as experience writing or customizing Kubernetes configurations