3 months ago
Santa Clara, CA, USAIntern
H1B Sponsor
Responsibilities
- Design, implement, and evaluate efficient deep neural network architectures for d-Matrix's AI compute engine.
- Collaborate with internal and external ML researchers to achieve R&D goals.
- Work with the Software team to meet stack development milestones.
- Conduct research to guide hardware design.
- Develop and maintain tools for high-level simulation and research.
- Port and optimize customer workloads for deployment and evaluate performance.
- Report and present progress effectively and timely.
- Contribute to publications and intellectual property.
Requirements
- Pursuing a Masters/PhD in Computer Science, Electrical and Computer Engineering, or a related field.
- High proficiency with PyTorch is essential.
- Strong skills in algorithm analysis, data structures, and Python programming are required.
- Current knowledge in machine learning and modern deep learning is necessary.
- Hands-on experience with modern neural network architectures like MoEs and Diffusion models is required.
- Knowledge of efficient deep learning techniques such as quantization and sparsity is preferred.
- Strong publication record in top machine learning conferences or journals is preferred.
- Proficiency in C/C++ programming is preferred.
- Experience with GPU CUDA programming is preferred.
- Familiarity with AutoML and meta learning is preferred.
- Experience with numerical analysis is preferred.
- Experience with specialized hardware accelerator systems for deep neural networks is preferred.