Description
We are seeking a Senior Solutions Architect with deep expertise in large-scale training and inference optimization of computer vision models such as diffusion-based.
Working alongside leading AI Native companies building the next generation of image, video, and multimodal experiences, you will serve as the hands-on technical bridge between NVIDIA's generative AI platform and the industry's most demanding production pipelines.
Your work will help customers unlock the full potential of NVIDIA's accelerated computing stack, driving measurable gains in performance, scalability, and efficiency across the full content generation lifecycle, from diffusion model training and optimization to image/video synthesis pipelines and multimodal foundation world models.
Responsibilities
- Guide EMEA AI Native companies building image, video, and multimodal generation products in training and deploying their pipelines on NVIDIA infrastructure.
- Provide deep technical guidance on diffusion model architectures (DiT, UNet, flow matching) and their efficient deployment across single and multi-GPU environments.
- Optimize generation pipelines for performance, scalability, and efficiency.
- Guide customers through the full visual content generation stack: codec-aware preprocessing, temporal consistency, video token representation, efficient long-video inference.
- Identify performance bottlenecks specific to vision workloads: memory-bound diffusion steps, attention scaling with resolution, and multi-GPU communication patterns for video.
- Translate customer's insights into actionable product feedback for NVIDIA's research and engineering teams.
- Contribute to the EMEA developer community through technical demos, workshops, and reference demos that showcase what is possible on NVIDIA stack.
Requirements
- MS or PhD in Computer Science, Computer Vision, Machine Learning, or equivalent hands-on experience.
- 5+ years in AI/ML with deep expertise Computer Vision models.
- Experienced with diffusion model frameworks for image/video generation.
- Understanding of vision encoder optimization, VAE architectures, and their performance tradeoffs at inference time.
- Strong communication skills, effective with ML researchers, creative technologists, and infrastructure engineers alike.
Nice to Have
- Familiarity with NVIDIA's inference stack: TensorRT, Triton Inference Server, and NIM.
- Hands-on experience on video generation: Temporal attention, 3D convolutions, or causal video transformers.
- Familiarity with codec-aware video pipelines and efficient video tokenization for generation at scale.
- Published work or benchmarks in image/video generation, diffusion acceleration, or visual foundation models.
NVIDIA offers highly competitive salaries and a comprehensive benefits package.