C++/PyTorch Developer Large Language Models

Job Description

Benefits:

Competitive salary

Opportunity for advancement

Training & development

Role: C++/PyTorch Developer Large Language Models

Location: Bay Area - CA (Hybrid)

Join one of our clients in the embedded systems space at the forefront of AI innovation. Were seeking a highly skilled C++ Developer with strong PyTorch expertise to help build and optimize Large Language Models (LLMs) that power next-gen machine intelligence.

Why Youll Love This Role:

Collaborate with world-class researchers and engineers building advanced LLMs.

Work at the intersection of high-performance C++ and cutting-edge Machine Learning.

Contribute to impactful AI deployments across real-world applications.What Youll Be Doing:

Design, develop, and optimize C++ code supporting Large Language Models using PyTorch.

Work directly with PyTorchs C++ API for integration and deployment.

Improve model efficiency in terms of performance, memory footprint, and scalability.

Debug and resolve issues across the stackC++, PyTorch, and hardware acceleration layers.

Engage in technical code reviews and maintain best practices for robust development.

Stay informed on the latest LLM architectures (e.g., LLaMA, BERT) and PyTorch features.

Collaborate in an Agile development setting with cross-functional ML teams.What Youll Need:

Minimum 3 years of hands-on C++ development experience.

Proficient in modern C++ (C++11/14/17), with strong coding standards.

Experience working with PyTorchs C++ APIs.

Familiarity with LLMs like BERT, GPT, or LLaMA.

Understanding of core ML concepts and neural network architectures.

Exposure to CUDA or other parallel programming tools.

Strong debugging skills and attention to technical detail.

Excellent communication and collaboration skills.

Experience in Linux/Unix-based development environments.Bonus Points:

Experience with TensorFlow, OpenCV, or similar frameworks.

Knowledge of Docker, Kubernetes, or other DevOps tools.

Flexible work from home options available.

Full-time

Hybrid remote