Research Engineer - Performance Optimization

Job Title:

Research Engineer - Performance Optimization

Position Type:

Full time

Location:

Palo Alto, CA, USA

Salary Range:

$180,000 - $250,000 (USD)

Job ID#:

156204

Job Description:

We are seeking engineers with substantial problem-solving experience in PyTorch, CUDA, and distributed systems. You will collaborate with Research Scientists to build and train cutting-edge foundation models on thousands of GPUs, focusing on multimodal generative models such as Diffusion Models and GANs. Experience in building inference or demo prototype code (including Gradio, Docker, etc.) is a plus.

Responsibilities:

• Ensure efficient implementation of models and systems for data processing, training, inference, and deployment.
• Identify and implement optimization techniques for massively parallel and distributed systems.
• Profile and optimize code to address bottlenecks in memory, speed, and utilization using high-performance CUDA, Triton, C++, and PyTorch.
• Collaborate closely with the research team to design efficient systems from start to finish.
• Develop tools for dataset visualization, evaluation, and filtering.
• Implement cutting-edge product prototypes based on multimodal generative AI.

Requirements:

• Experience training large models with Python & PyTorch, covering the entire development pipeline from data processing to inference.
• Proficiency in optimizing and deploying inference workloads for throughput and latency.
• Experience profiling CPU & GPU code in PyTorch, using tools like Nvidia Nsight.
• Skill in writing and improving highly parallel and distributed PyTorch code, with familiarity in DDP, FSDP, Tensor Parallel.
• High-performance C++ coding experience, preferably within an ML context.
• Expertise in high-performance Triton / CUDA programming and custom PyTorch kernel development, including utilizing tensor cores.
• Knowledge of deep learning concepts such as Transformers, Diffusion Models, and GANs is advantageous.
• Experience building inference/demo prototypes with tools like Gradio and Docker is beneficial.

About Us:

Founded in 2009, IntelliPro is a global leader in talent acquisition and HR solutions, operating in over 160 countries including the USA, China, Canada, Singapore, Japan, Philippines, UK, India, Netherlands, and the EU. We are committed to diversity and inclusivity, valuing all candidates regardless of background.

Learn more at .

Compensation:

The offered salary will depend on various factors such as education, experience, location, and certifications. We also provide a comprehensive benefits package, subject to eligibility.

#J-18808-Ljbffr