10 months ago
Remote, United States +3 moreSenior
Responsibilities
- Design and implement improvements to large-scale model training infrastructure.
- Contribute to technical decisions that optimize model performance.
- Work on post-training processes, including reinforcement learning and fine-tuning.
- Improve the efficiency and scalability of model serving infrastructure.
- Profile GPU-accelerated workloads, identify bottlenecks, and implement performance tuning techniques.
- Scale GPU jobs across large compute clusters.
Requirements
- Strong engineering skills with fluency in Python and PyTorch or other frameworks.
- Proven experience implementing and training large deep learning models.
- Experience writing and debugging low-level GPU code with CUDA and C++.
- Experience scaling GPU jobs using large-scale compute clusters such as Slurm or Kubernetes.
- Demonstrated ability to analyze and optimize GPU-accelerated workloads.
Benefits
- Remote-first work environment with a globally distributed team.
- Five weeks of paid leave.
- Comprehensive healthcare benefits, including vision and dental.
- Additional well-being perks.
- Visa assistance, including H1B and OPT transfers, for US employees.
Tech Stack
Categories
About Reka
Reka is a privately held AI research and product company building multimodal foundation models and tools for organizations and businesses. Headquartered in Sunnyvale with a remote-first team, it offers products such as Reka Vision and Reka Edge, a 7B vision-language model. Founded by researchers from Google DeepMind and Meta’s FAIR, it also operates Claru, a data unit supplying licensed training datasets for physical and embodied AI.