Senior Machine Learning Engineer, Cosmos Engineering

NVIDIA•Santa Clara, CA

7d•$184,000 - $356,500

About The Position

NVIDIA’s Metropolis team builds intelligent video analytics and perception solutions for smart cities, industrial automation, and autonomous systems. The Senior Machine Learning Engineer will develop production-quality Metropolis AI capabilities powered by Cosmos, NVIDIA’s world foundation model platform, addressing real-world perception and physical AI challenges. The role also partners with the Cosmos team to expand platform features in alignment with the Metropolis product roadmap. The engineer will identify model capability gaps and recommend solutions, including synthetic data generation, improved tuning methods, and architecture enhancements. The role leads open-sourcing of team solutions and research artifacts, keeps current with advances in foundation models and generative AI, and collaborates across Product, Program, Engineering, and Data Procurement teams to drive alignment and execution.

Requirements

MSc or PhD in Computer Science, Electrical Engineering, or a related field, or equivalent experience.
8+ years of experience in applied machine learning or AI research.
Deep expertise in deep learning fundamentals and hands-on experience with diffusion models and generative architectures.
Experience pre-training or refining large language models (LLMs), vision-language models (VLMs), or world foundation models (WFMs).
Experience with large-scale foundation models, including training workflows, fine-tuning techniques, and evaluation approaches.
Experience with simulation environments such as Isaac Sim or similar platforms.
Track record of leading projects end-to-end and delivering results with clarity and accountability.
Ability to manage and complete tasks across multiple parallel workstreams in a fast-paced, evolving environment.
End-to-end understanding of the ML development and deployment lifecycle, and ability to adopt and use modern AI development tools and workflows.

Nice To Haves

Hands-on experience with model compression, quantization, and real-time inference optimization for production deployments.
Experience implementing AI solutions in physical settings such as public areas, smart infrastructure, or robotics platforms.
Experience scaling AI systems across distributed infrastructure, including multi-node training and large-scale data pipelines.
Published research or open-source contributions in generative models, synthetic data, or physical AI.
Familiarity with CUDA, Triton, or low-level GPU kernel development for inference pipeline acceleration.

Responsibilities

Build and deliver Metropolis AI solutions powered by Cosmos world foundation models for intelligent video analytics, perception, and physical AI.
Identify areas where Cosmos models underperform or lack capabilities, and recommend solutions such as synthetic data generation, refinement of tuning methods, and architecture improvements.
Partner with the Cosmos team to identify and prioritize platform features aligned with the Metropolis product roadmap and contribute deliverables supporting the wider ecosystem.
Lead open-sourcing of team-developed solutions and research artifacts.
Stay current with advances in foundation models, generative architectures, and training methodologies, and bring relevant insights to the team.
Collaborate with Product, Program, Engineering, and Data Procurement teams to drive alignment and unblock execution.