Member of Technical Staff (GPU Performance Engineer)
Reka · Remote · US, UK, Singapore, Remote ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- Fully remote (Global)
- Python + PyTorch
- CUDA + C++
- Visa sponsorship (US)
- Senior level
Join Reka as a Member of Technical Staff focused on GPU performance. You will design improvements for training infrastructure, optimize model performance, and enhance the efficiency of model serving systems. The team works on state-of-the-art multimodal artificial intelligence, leveraging cutting-edge hardware and software to push the boundaries of AI innovation.
What you'll do
- Design and implement improvements to training infrastructure
- Optimize performance of large-scale deep learning models
- Work on post-training processes like reinforcement learning
- Improve efficiency and scalability of model serving infrastructure
- Make technical decisions to enhance system performance
What you'll need
- Strong engineering skills with fluency in Python
- Experience with PyTorch or similar frameworks
- Proven track record in training large deep learning models
- Ability to write and debug low-level GPU code (CUDA, C++)
- Experience scaling GPU jobs on clusters (Slurm, Kubernetes)
- Skill in profiling and optimizing GPU-accelerated workloads
Stack
- Python
- PyTorch
- CUDA
- C++
- Slurm
- Kubernetes
How you'll work
- Remote-first company
- Globally distributed team
- Headquartered in San Francisco Bay Area
What you get
- Five weeks of paid leave
- Comprehensive healthcare benefits
- Vision and dental coverage
- Visa assistance for US employees (H1B, OPT)