Senior Software Engineer-Distributed Inference
NVIDIA- Full Time
- Senior (5 to 8 years)
Candidates should have a Bachelor's degree with a minimum of 6+ years of professional experience in software development focused on C++, or a Master's degree preferred in computer science, engineering, or a related field with 3+ years of relevant experience. A strong background in architecting and building complex software systems is required, along with experience in distributed systems or high-performance computing (HPC) applications. Familiarity with PyTorch internals or similar machine learning frameworks is a significant advantage. Technical skills should include strong proficiency in modern C++ (C++11 and above) and Python, a solid understanding of software design patterns and best practices, experience with parallel and concurrent programming, and proficiency in CMake and Pytest. Knowledge of GPU programming and acceleration techniques is a plus.
The Staff Software Engineer will lead the design and implementation of a high-performance inference runtime that leverages d-Matrix's advanced hardware capabilities. They will integrate the inference runtime with PyTorch to enable upstream software capabilities like inference and finetuning. The role involves collaborating closely with cross-functional teams including hardware engineers, data scientists, and product managers to define requirements and deliver integrated solutions. Additionally, they will develop and implement optimization techniques to ensure low latency and high throughput in distributed and HPC environments, ensure code quality and performance through rigorous testing and code reviews, and create technical documentation to support development, deployment, and maintenance activities.
AI compute platform for datacenters
d-Matrix focuses on improving the efficiency of AI computing for large datacenter customers. Its main product is the digital in-memory compute (DIMC) engine, which combines computing capabilities directly within programmable memory. This design helps reduce power consumption and enhances data processing speed while ensuring accuracy. d-Matrix differentiates itself from competitors by offering a modular and scalable approach, utilizing low-power chiplets that can be tailored for different applications. The company's goal is to provide high-performance, energy-efficient AI inference solutions to large-scale datacenter operators.