WILL is an entrepreneurial laboratory incubated by WizardQuant. Starting from but not limited to financial scenarios, WILL sincerely invites students, new graduates, and experienced professionals to join us and embark on a journey across vast expanse of artificial intelligence.

AI engineers build and optimize infrastructure and software stacks for ASI for Sci-Tech, and are deeply involved in the co-design of algorithms and systems for LLMs. Your responsibilities include but are not limited to computing cluster optimization, LLM training pipelines, efficient training and inference, and MLOps.

We are looking for candidates who are proficient in modern AI tool chains and large model infrastructure, and who are passionate about implementing the outcomes of cutting-edge research, with the goal of building the technical foundation for advanced AI systems and ensuring their efficient operation.

Core Responsibilities

• Computing cluster optimization: Participate in the construction of AI infrastructure, optimize network, storage, and hardware, build a solid and efficient AI foundation.
• Software platform construction: Design a high-performance and scalable AI application service architecture, ensure high availability and low latency in the production environment.
• Training pipeline optimization: Quicklyiterate LLM training pipelines, including data preprocessing and distributed training, ensure that the training process is robust and scalable.
• Model optimization: Collaborate with AI researchers to continuously optimize model training and inference Carry outperformance analysis and tuning to resolve system bottlenecks.

Qualifications

• Education background incomputer science, engineering or other related fields, with solid CS skills.
• Proficient in Python and C++.
• Expertise in high-performance computing, distributed computing, MLOps and deployment, model optimization, big data systems, etc
• Passionate about implementing cutting-edge AI solutions, self-motivated,collaborative mindset, and able to adapt to fast-paced challenges.
• Technical expertise or implementation experience in the following areas is a plus:
• 1) Familiar with RDMA protocol,knowledgeable in distributed communication primitives optimization;
• 2) Familiar with general GPU architecture, with CUDA and CUDA-like operator performance optimization experience.