High-performance computing development (including machine learning platform/deep learning framework/operators, etc.) high salary no rolling

By the end of 2025, there will be 55 billion-dollar quantitative private equity firms in the country, of which 47 will be our customers; there will also be more than 100 quantitative private equity firms with 0 to 100 billion in assets, 15 (non-asset management) quantitative self-managed funds, and a small number of foreign quantitative hedge funds, all of which will closely cooperate with our company.

[We will not charge you any fees!]

Please note:

1. This position does not require immediate employment, and we welcome inquiries at any time!

3. If you are currently working in the same industry or have already received an offer from a certain company, I will recommend anonymously from your current company and hide your name, and supplement appropriate privacy protection measures until you initiate an interview or make significant progress.

[The following is a formal introduction]

We provide you with more than just a job, but also a chance to work with a top-notch team:

2. Stable work and life balance (usually 965), and leave plenty of free space within working hours for you to deepen your learning, explore your interests, focus your thoughts or relax and rejuvenate.
3. With a pleasant rest area and self-run coffee shop, drinks and snacks are supplied around the clock; equipped with an employee fitness center, high-end ergonomic chairs and high-definition multi-screen workstations, we create an ideal office environment in every way.
5. The team is composed of elite from renowned universities and industry experts. Rich research directions are available for selection, encouraging personalized development; technical sharing is held regularly every week, with an open atmosphere and harmonious collaboration.
6. In addition to employee funds, zero-cost dividends, partner plans, quarterly special contributions, and generous year-end bonuses, we place greater emphasis on sharing every growth joy with you.

Duties:

1. Responsible for the architecture design and development of high-performance computing and machine learning platforms, building a unified low-latency online inference and high-performance transaction execution infrastructure that supports large-scale distributed training, and continuously optimizing the system.
3. Deeply optimize training and inference performance, solve bottlenecks from data loading, computational pipeline, operator execution to distributed communication, and improve the utilization of GPUs and other heterogeneous hardware and overall computing efficiency.
5. Participate in the construction and management of high-performance computing clusters, involving Linux kernel, high-speed network (such as RDMA/DPDK), resource scheduler (such as Slurm/Kubernetes) configuration, optimization and secondary development.
6. Responsible or involved in optimizing the compiler and low-level toolchain, targeting specific hardware (CPU/GPU/NPU) for compilation optimization, instruction generation, or using MLIR/TVM/XLA tools to improve the execution efficiency of the computation graph.
7. Track and apply cutting-edge technologies such as cloud-native computing, heterogeneous computing, and compilation techniques in these fields, and promote platform advancedness and rapid business requirement adaptation in terms of technology.
8. Collaborate with quantitative researchers and algorithm engineers to efficiently convert strategy models into high-performance production code, solving performance, stability, and deployment challenges throughout the research-to-trading pipeline.

Requirements:

1. Proficient in at least one language such as C++/Golang/Python, with solid foundations in computer architecture, data structures and algorithms, and system-level debugging and optimization abilities.
2. In-depth understanding of distributed system principles, familiar with cloud-native technology stack (Kubernetes/Docker), network programming, databases and monitoring systems, and large-scale system development experience.
3. Familiar with mainstream deep learning frameworks (TensorFlow/PyTorch) at the bottom level mechanism and architecture, familiar with MLOps concept and related tools (such as Kubeflow, MLflow), and have large-scale training/inference system optimization experience.
6. Familiar with the resource management, performance analysis and optimization methods of heterogeneous computing hardware (GPU/FPGA, etc.), with relevant development or tuning experience.
7. Understand the common processes and challenges of quantitative trading and financial data processing, or those with relevant system development experience in the field are preferred.