Drimva

NVIDIA

Solutions Architect, LLM Model Builder

$152K–$242K

US, CA, Santa Clara · Full-time · Posted Apr 7, 2026

Experience
5+ years of relevant experience; Research Engineer, Deep Learning Engineer, or equivalent.
Education
MSc, PhD in Computer Science, Electrical Engineering, Software Engineer, ML Engineer, or related fields (or equivalent experience).

About the job

Advises partners on building, benchmarking, optimizing, and deploying foundation-model solutions and production inference for customer workloads.

What you'll do

  • Guide partners on model fine-tuning, optimization, benchmarking, evaluation, and inference architecture.
  • Define benchmark plans, evaluation workflows, compute plans, and production-readiness testing.
  • Develop reference architectures and technical resources, and support pre- and post-sales partner engagements.

What they ask for

  • 5+ years of relevant experience with LLMs, VLMs, and large-scale inference systems, including hands-on model development and deployment.
  • Strong foundation-model knowledge and familiarity with reasoning models, reinforcement learning, and synthetic-data workflows.
  • Strong programming skills in Python and hands-on experience with PyTorch, JAX, or TensorFlow.
  • Strong communication and presentation skills for technical teams and executives.

Nice to have

  • Experience deploying large-scale AI systems for partners or customers; building benchmark suites, fine-tuning recipes, sizing calculators, or TCO models.
  • Knowledge of GPU infrastructure and cluster technologies; active open-source contributions in model tooling, inference, evaluation, or performance optimization.

Pay

152,000 USD - 241,500 USD (base salary)

Tools

Python, PyTorch, JAX, TensorFlow, Nemotron, NeMo, Dynamo, TensorRT-LLM

AI focus

Partner enablement for reasoning and multimodal models, fine-tuning, synthetic data, evaluation, optimization, and production inference.

Model-written summary · the original posting is authoritative

First seen Oct 15, 2026 · Last verified Nov 5, 2026 (UTC)