Solutions Architect, AI Models

NVIDIA NVIDIA · Semiconductors · Santa Clara, CA +1 · Remote

NVIDIA is seeking a Solutions Architect to help enterprise customers adopt NVIDIA AI software and models. This role involves developing end-to-end AI solutions, tackling complex challenges across the AI model lifecycle (data processing, orchestration, training, post-training, RL, evaluation, optimization), and supporting a broad model portfolio. The architect will partner with customers to understand their needs and deliver customized AI solutions, contributing to product improvement and sharing knowledge through open-source projects, product engineering, or training.

What you'd actually do

  1. A huge part of our work involves developing end-to-end AI solutions for enterprise use cases. We help customers adopt NVIDIA AI models and libraries by offering deep technical expertise.
  2. Tackle sophisticated AI challenges by applying skills across the AI model lifecycle—from data processing and orchestration to training, post-training, reinforcement learning (RL), evaluation, and model optimization.
  3. Support a broad model portfolio spanning LLMs, multimodal, retrieval, speech, content safety, and edge use cases.
  4. Partner with enterprise customers in co-design engagements — understanding their data, evaluation criteria, and success metrics to deliver customized AI solutions.
  5. As we work with customers across multiple industries, we help improve NVIDIA products and build creative solutions to overcome scaling challenges at the intersection of computer architecture, models, libraries, and AI applications.

Skills

Required

  • BS, MS, or Ph.D. degree in Engineering, Mathematics, Physics, Computer Science, Data Science, or similar (or equivalent experience)
  • 5+ years of experience with AI frameworks such as PyTorch, JAX, or TensorFlow, and libraries like Hugging Face Transformers.
  • Proficiency in Python programming, software design, debugging, and performance analysis, with at least 5+ years of experience in a Linux environment.
  • Hands-on experience with AI model lifecycle, including evaluation, failure analysis, pre-training, post-training, RL, and model optimization.
  • Expertise in distributed computing methodologies, including model and data parallelism.
  • Experience with distributed computing tools, like SLURM and Kubernetes, for training and evaluating large models on GPUs.
  • Ability to learn fast and quickly adapt to change.
  • Clear written and oral communications skills with the ability to effectively collaborate with executives and engineering teams.

Nice to have

  • Experience with and/or contributions to open-source NVIDIA AI libraries and models, particularly Nemotron, NeMo, NeMo Framework, NeMo-RL.
  • Hands-on experience with data curation and analysis for model post-training and RL.
  • Prior experience with AI model training techniques applied to multi-modal data (audio, image, and video).
  • Knowledge of NVIDIA GPU/CPU architecture and its impact on software performance.
  • Willingness and ability to dig into unfamiliar territories to solve complex problems relying on experience from previous work.

What the JD emphasized

  • enterprise AI solutions
  • AI model lifecycle
  • training
  • post-training
  • reinforcement learning (RL)
  • evaluation
  • model optimization
  • distributed computing methodologies
  • model and data parallelism
  • distributed computing tools
  • large models on GPUs

Other signals

  • customer-facing role
  • deploying AI workloads at scale
  • enterprise AI solutions
  • NVIDIA AI software adoption