Software Engineer - AI Agent (Chinese speaking)

re-zoo-me; Northern, New York (KY, NY); Hybrid; USD 130,000 - 190,000; Full time

Job summary

re-zoo-me seeks a seasoned engineer to shape and scale our machine learning platform across data processing, training, and live deployment for trading systems.

You will design data pipelines, optimize training/inference infra, and implement experiment tracking and orchestration to improve iteration speed and model trust in production. Collaboration and ownership of performance bottlenecks are essential.

Qualifications

• Solid foundation in computer engineering with the ability to understand and analyze complex systems.
• Hands-on experience in one or more areas: Large-scale data systems, Distributed systems, ML platforms, Model training and inference, or System performance optimization/GPU optimization.
• Ability to debug and iteratively improve systems under real-world constraints.
• Strong sense of ownership: proactively identifying and defining problems.

Responsibilities

• Design and develop data processing systems transforming high-throughput trading data into stable, production-ready model inputs.
• Build and optimize model training and inference infrastructure.
• Develop experimentation and evaluation systems to improve iteration efficiency and trustworthiness of results.
• Improve the end-to-end workflow of the ML platform for live trading deployment.
• Identify and resolve performance bottlenecks across compute, memory, scheduling, and I/O.

Skills

Large-scale data

Distributed systems

ML platforms

Training/inference

Performance optimization

GPU optimization

Kubernetes

Ray

Kubeflow

Megatron-LM

DeepSpeed

vLLM

Tools

Megatron-LM

DeepSpeed

vLLM

Kubernetes

Ray

Kubeflow

Volcano

Job description

Key Responsibilities:

• You will play a critical role in key system components spanning the entire lifecycle of machine learning models—from research to live trading deployment. Depending on your technical expertise, you will contribute to one or more of the following areas:
• Design and develop data processing systems that transform high-throughput, high-noise trading data into stable, efficient, and production-ready model inputs.
• Build and optimize model training and inference infrastructure.
• Develop experimentation and evaluation systems to improve model iteration efficiency, while enhancing traceability and trustworthiness of experimental results.
• Continuously improve the end-to-end workflow of the machine learning platform, enabling stable deployment and operation of models in live trading environments.
• Identify and resolve performance bottlenecks across the entire technology stack, conducting systematic optimizations in areas such as compute, memory, scheduling, and I/O.

Requirements:

• Solid foundation in computer engineering and the ability to understand and analyze complex systems.
• Hands-on experience in one or more of the following areas:
• 1. Large-scale data systems
• 2. Distributed systems
• 3. Machine learning platforms
• 4. Model training and inference systems
• 5. System performance optimization or GPU performance optimization
• Ability to debug and iteratively improve systems under real-world constraints.
• Strong sense of ownership: not only executing assigned tasks with high quality but also proactively identifying and defining problems.

Plus:

• Solid knowledge and hands-on experience in performance optimization for accelerators such as GPUs, TPUs, as well as memory and communication subsystems.
• Familiarity with large-scale machine learning training and inference systems, with hands-on experience in mainstream frameworks such as Megatron-LM, DeepSpeed, or vLLM.
• Experience in building machine learning infrastructure, including key components such as experiment tracking, workflow orchestration, and model serving.
• Familiarity with the Kubernetes ecosystem, or experience developing or maintaining distributed scheduling systems such as Ray, Kubeflow, or Volcano.