Role Description:
The engineer will be responsible for the design and development of a computer vision compiler targeting neural networks, transformers, and large language models (LLMs). This role involves leveraging a variety of techniques, including operation expansion, load balancing, data pipelining, and advanced optimization strategies to improve efficiency and performance throughout the compilation process.
Key responsibilities:
1. Develop an optimization compiler for computer vision models and LLMs.
2. Perform hardware-aware compiler optimizations.
3. Research methods to improve compilation time, inference speed, and model accuracy.
4. Provide technical support and solutions for model optimization and deployment.
Requirements:
1. Education background: Master degree or Ph.D degree
2. Minimum experience: At least one year of relevant work or academic experience
3. Similar or other experiences:
- Familiarity with computer architecture and parallel computing is a plus.
- Experience with GDB debugging is a plus.
- Background in machine learning, computer vision, or LLMs is a plus.
4. Skills
- Proficiency in C/C++ programming is required.
- Familiarity with Python and the Linux environment.
- Ability to understand and interpret hardware specification documents.