Senior Software Engineer - VLM Microservices for Neural Reconstruction

at Nvidia

USD 152,000-287,500 per year

SENIOR

✅ On-site

Tech Stack
Tag name is followed by "@" symbol and proficiency level value. About proficiency levels:

1-2 — basic awareness. Minimal hands-on experience, and a rudimentary understanding of the technology's purpose;

3-6 — daily use. Comfortable and regular usage, capable of handling common tasks and challenges related to the technology;

7-9 — you are an expert, you can teach others, you know all the pitfalls and tricks;

10 — exceptional knowledge, comprehensive understanding, and adeptness in all aspects of the technology, including advanced problem-solving. Think twice before claiming or demanding such level.

AI @ 6 API @ 6 CI/CD @ 7 CUDA @ 4 Communication @ 6 Computer Vision @ 7 Distributed Systems @ 6 Docker @ 6 Helm @ 6 Kubernetes @ 6 LLM @ 4 Machine Learning @ 4 Microservices @ 6 Python @ 6 Robotics SGLang @ 4 Security gRPC @ 6 vLLM @ 4

Details

NVIDIA's team builds the Omniverse NuRec SDK, enabling robotics, healthcare, and autonomous vehicle developers to create better models faster through closed-loop validation and training grounded in real-world scenarios.

The role focuses on bringing Vision Language Models and 3D reconstruction technologies into NVIDIA's neural graphics software ecosystem to improve robustness, accuracy, and capabilities, helping bridge the gap between the real world and simulations.

Responsibilities

Design, build, and optimize containerized inference execution for NVIDIA's latest 3D Vision Language Models, turning research work into production-grade, highly optimized software, including NVIDIA Inference Microservices (NIMs).
Develop benchmarks to validate model accuracy and performance, including latency, throughput, and scalability.
Release and maintain models and their pipelines throughout their lifecycle, including bug fixes and security patches.
Contribute Vision Language Model features to open-source projects such as vLLM.
Collaborate closely with Research and Product teams and influence shared roadmaps.

Requirements

Master's degree in Computer Science with 3 years of experience, or a bachelor's degree in Electrical Engineering or equivalent with 5 years of experience.
History of building, validating, and releasing production-grade AI distributed systems, backend services, microservices, and cloud technologies.
Deep technical expertise in distributed applications using Docker, Kubernetes, endpoints and APIs such as REST and gRPC, and Helm.
Hands-on experience with modern inference platforms, including vLLM, SGLang, Torch, TRT, and TRT-LLM.
Proficiency in Python and C++.
Strong software engineering fundamentals, including source control, CI/CD, testing and validation, packaging, and containerization.
Excellent written, visual, and verbal communication skills.
Curiosity and willingness to learn new technologies and collaborate across teams and functions.

Preferred Qualifications

Track record of contributing to open-source or production-grade software, particularly vLLM.
Experience with machine learning model engineering, including training, fine-tuning, distillation, and quantization.
Experience with low-level machine learning model optimization, including CUDA kernels.
Strong fundamentals in 3D graphics, 3D computer vision, or neural reconstruction, including NeRFs and Gaussian Splatting.
History of multidisciplinary creativity and innovation across multiple software engineering domains.

Benefits

Equity and employee benefits are provided.
NVIDIA is an equal opportunity employer committed to fostering a diverse work environment.

The base salary range is USD 152,000–241,500 for Level 3 and USD 184,000–287,500 for Level 4. Applications will be accepted at least until May 2, 2026.