Senior Research Engineer, Interactive World Models at NVIDIA

Department: Engineer, Sys SW

Compensation

$184,000 – $356,500/yr

Setup: On-site
Location: California
Type: Full-time
Level: senior
Posted: Aug 20, 2026

Description

NVIDIA is hiring a Senior Research Engineer to lead engineering across model development and runtime systems for its FlashDreams and FastGen world-model technologies. The role focuses on optimizing continuous autoregressive serving, GPU inference, KV-cache and frame-streaming management, few-step distillation, causal or autoregressive generation, and end-to-end delivery of multi-user simulation workflows. Candidates need an MS or PhD in a relevant field, at least five years of applied machine learning or research engineering experience, and hands-on experience with Python, PyTorch, GPU training and inference, and performance optimization. The posting states that applications were accepted at least until August 24, 2026, and provides base salary ranges for Level 4 and Level 5 engineers.

Required Skills

Python
PyTorch
GPU inference
KV-cache management
Model distillation
Real-time ML systems

Additional Information

Minimum Education: Master's or PhD in Computer Science, Electrical Engineering, or a related field, or equivalent experience
Required Previous Experiences: At least 5 years of equivalent experience in applied machine learning or research engineering
Experience in video or world models, generative modeling, model distillation, simulation, robotics, computer vision, or real-time stateful machine learning systems
Hands-on experience with Python, PyTorch, GPU-accelerated training and inference, performance optimization, or serving
Preferred Qualifications: Experience with post-training generative video models, including distillation, self-forcing, action conditioning, or long-horizon memory
Experience building and optimizing real-time, stateful generative inference systems
Technical stewardship of an open-source machine learning project