Machine Learning Engineer at IMC | Chicago, Hong Kong, Sydney

Your Core Responsibilities: Design and implement large-scale distributed training pipelines to efficiently manage extensive datasets and intricate models. Optimize low-latency inference pipelines to ensure production systems deliver real-time predictions. Create libraries that enhance the performance of existing machine learning frameworks. Leverage GPU hardware and acceleration libraries to maximize performance in both training and inference stages. Architect scalable model frameworks capable of processing high-volume trading data, ensuring high accuracy in real-time predictions. Work alongside quantitative researchers to automate machine learning experiments, hyperparameter tuning, and model retraining processes. Collaborate with high-performance computing specialists to refine workflows, enhance training speeds, and minimize costs. Assess and integrate third-party tools to boost model development, training, and inference capabilities. Deep dive into the functionalities of open-source machine learning tools to extend their capabilities and improve overall performance. Your Skills and Experience: 5+ years of experience in machine learning, particularly in training or inference systems. Hands-on experience with low-latency, real-time machine learning pipelines in high-performance settings is highly desirable. Strong programming skills, particularly in Python, CUDA, or C++. Familiarity with machine learning frameworks such as PyTorch, TensorFlow, or JAX. Proficient in GPU programming for both training and inference acceleration (e.g., CuDNN, TensorRT). Experience with distributed training techniques for scaling machine learning workloads (e.g., Horovod, NCCL). Exposure to cloud platforms and orchestration tools is beneficial. A proven record of contributing to open-source projects in machine learning, data science, or distributed systems is a plus.