Senior Machine Learning Engineer - Physical AI and Synthetic Data Generation and Evaluation

NVIDIA Gruppe; Santa Clara (CA); On-site; USD 224,000 - 431,250; Full time

Benefits offered by this job

Equity

Benefits

Job summary

NVIDIA in Santa Clara invites outstanding Machine Learning Engineers to join our Physical AI teams, driving high-fidelity synthetic data for autonomous vehicle simulations and QA tooling.

You will architect Generative pipelines, train multimodal models, and lead data quality validation across complex synthetic environments, collaborating across AI, software, and hardware domains to push the frontier of real-time AI in the physical world.

Qualifications

• BS, MS or PhD in Computer Science, Computer Graphics, Robotics, or related field (or equivalent experience).
• 12+ years of experience in ML software development.
• Deep technical knowledge of image/video synthesis, including diffusion models and state-of-the-art multimodal methods.
• Strong hands-on skills in major DNN libraries and Python among others. Experience with workflow management and databases for large-scale training and data generation.
• Strong analytical and mathematical skills bridging data-driven approaches and physical world constraints.
• Collaborative outlook with outstanding communication skills, thriving in a tightly-knit team.
• Experience in assessing the impact of synthetic data on model performance through metrics and systematic validation.

Responsibilities

• Architect Generative Pipelines: Develop and implement advanced image and video generation/editing/reasoning models to produce high-fidelity synthetic data for Physical AI applications.
• Multimodal Development: Build and fine-tune large-scale models, including VLMs, MLLMs, Generation models, applying transformer, auto-regressive and diffusion-based architectures.
• Controllable Synthesis: Apply and evolve user controls during data generation to ensure precise environmental and structural control over generated data.
• Automated Quality Assurance for Sensor Data and Ego Policy: Build and test automated data QA pipeline using MLLMs and a mix of well known classical algorithms.
• Detailed Validation: Establish KPI evaluation and validation to ensure quality and physical accuracy of synthetic releases; design benchmark datasets.
• SOTA Data Engineering: Lead the generation of massive training datasets using various state-of-the-art tools and synthetic data mining techniques.
• Contribute to the full lifecycle of ML software, including performance optimization, testing, and high-quality documentation.

Skills

Python

DNN libraries

ML software development

Multimodal ML

Diffusion models

Education

BS/MS/PhD in Computer Science, Computer Graphics, Robotics, or related field

Job description

We are looking for outstanding Machine Learning Engineers to join our Physical AI teams. As the pioneers of the GPU - the visual cortex of modern computing, we are building the foundation for the next wave of AI that interacts with the physical world.

This role is at the forefront of Physical AI, developing sophisticated reasoning modules to build high-fidelity synthetic datasets. It leverages state-of-the-art multimodal models and diffusion techniques to simulate complex physical environments, ensuring our AI agents are trained on the most diverse and rigorous data possible. In particular, we will build advanced quality assurance technology to validate generated data outputs. We work closely with various users of synthetic datasets, including policy models. It extends an opportunity to contribute to the technology that will drive the cars of the future!

What you’ll be doing:

• Architect Generative Pipelines: Develop and implement advanced image and video generation/editing/reasoning models to produce high-fidelity synthetic data for Physical AI applications.
• Multimodal Development: Build and fine-tune large-scale models, including VLMs, MLLMs, Generation models, applying transformer, auto-regressive and diffusion-based architectures. These models will take both visual and structured inputs (such as world model representations), and generate data and analyze consistency between scenarios intended by users and the generated data.
• Controllable Synthesis: Apply and evolve user controls during data generation to ensure precise environmental and structural control over generated data.
• Automated Quality Assurance for Sensor Data and Ego Policy: Build and test automated data QA pipeline using MLLMs and a mix of well known classical algorithms. In particular, build new capabilities to judge the quality of behavioral policies to ensure high quality data delivery for VLA.
• Detailed Validation: Establish a strong mentality for KPI evaluation and validation to ensure the quality and physical accuracy of the synthetic releases. Establish a benchmark dataset. Design and validate KPI metric designs.
• SOTA Data Engineering: Lead the generation of massive training datasets using various state-of-the-art tools and synthetic data mining techniques.
• Contribute to the full lifecycle of ML software, including performance optimization, testing, and high-quality documentation.

What we need to see:

• BS, MS, or PhD in Computer Science, Computer Graphics, Robotics, or a related field (or equivalent experience).
• 12+ years of experience in ML software development.
• Deep technical knowledge of image/video synthesis, including diffusion models and state-of-the-art multimodal methods.
• Strong hands-on skills in major DNN libraries and computer languages including Python among others. Various hands on experience with workflow management and database to facilitate large scale training and data generation. Strong skills to optimize code efficiency is a huge plus.
• Strong analytical and mathematical skills to bridge the gap between data-driven approaches and physical world constraints.
• A collaborative outlook with outstanding communication skills, thriving in a tightly-knit team environment.
• Experience in assessing the impact of synthetic data on model performance through metrics and systematic validation.

Ways to stand out from the crowd:

• Experience with computer/GPU architecture to improve the performance during inference/training.
• Familiarity with simulation platforms and deep understanding of 3D sensor modalities (Camera, Multi cameras, Lidar, Radar).
• Experience with open source software.

We value different paths to technical excellence and welcome candidates who bring strong judgment, curiosity, and a collaborative approach. Come build the future of autonomous vehicle simulation with us!

Your base salary will be determined based on your location, experience, and the pay of employees in similar positions. The base salary range is 224,000 USD - 356,500 USD for Level 5, and 272,000 USD - 431,250 USD for Level 6.

You will also be eligible for equity and benefits.

This posting is for an existing vacancy.

NVIDIA uses AI tools in its recruiting processes.

NVIDIA is committed to fostering an inclusive work environment and proud to be an equal opportunity employer.

As we highly value diversity in our current and future employees, we do not discriminate (including in our hiring and promotion practices) on the basis of race, religion, color, national origin, gender, gender expression, sexual orientation, age, marital status, veteran status, disability status or any other characteristic protected by law.