Senior Multimodal AI Solutions Architect

NVIDIA; Marseille; Sur place; EUR 120 000 - 180 000; Plein temps

Résumé du poste

NVIDIA in Marseille is seeking a Senior Solutions Architect to drive multimodal AI solutions for document intelligence, training models and optimizing production deployments across image, video and text modalities.

You will partner with NVIDIA research and engineering to translate field insights into roadmaps for NeMo, TensorRT-LLM, Dynamo and RAPIDS, while engaging customers with technical talks, demos and reference blueprints.

Qualifications

• MS or PhD in Computer Science, Engineering, or equivalent.
• 7+ years in applied AI/ML focused on document understanding and visual content analysis.
• Proven track record in building or optimizing VLMs and Omni models.
• Familiarity with NVIDIA ecosystem: TensorRT-LLM, NeMo, RAPIDS.

Responsabilités

• Develop technical relationships with customers building multimodal AI systems for document intelligence and image/video analysis, guiding them from architectural planning through deployment.
• Advise on model training strategies across modalities including layout-aware encoders, user-item interaction models, and spatiotemporal video representations.
• Optimize vision encoders for production latency and efficient video sampling and temporal reasoning.
• Represent customer needs to NVIDIA product teams, contributing insights to roadmaps for NeMo, TensorRT-LLM, Dynamo and RAPIDS.
• Engage the developer community through hackathons, technical talks, demos, and reference blueprints.

Connaissances

Document understanding

Multimodal ML

Model optimization

Communication

Formation

MS or PhD in Computer Science or Engineering

Outils

TensorRT-LLM

NeMo

RAPIDS

Description du poste

NVIDIA in Marseille is seeking a Senior Solutions Architect to drive multimodal AI solutions for document intelligence, training models and optimizing production deployments across image, video and text modalities.

You will partner with NVIDIA research and engineering to translate field insights into roadmaps for NeMo, TensorRT-LLM, Dynamo and RAPIDS, while engaging customers with technical talks, demos and reference blueprints.