AI Research Engineer - Finance-Driven ML Innovations

P2P; New York (NY); On-site; USD 255,000 - 345,000; Full time

Job summary

Jump Trading Group seeks research scientists to apply machine learning to state-of-the-art capabilities across complex domains. You will drive end-to-end research from concept to production, iterating on model design, tooling, and infrastructure.

You will collaborate with researchers and quants, optimize training pipelines on HPC resources, and deploy models where latency matters, using Python/C++ and frameworks like PyTorch, JAX, or TensorFlow.

Qualifications

• Strong publication record at ICML, ICLR, AAAI, NeurIPS, UAI, KDD, or equivalent.
• Strong general ML background with deep learning or language modeling exposure.
• Solid development skills in Python and/or C++.
• Familiarity with ML libraries/frameworks such as PyTorch, JAX, and/or TensorFlow.
• Collaborative, pragmatic mindset with ability to communicate with researchers.

Responsibilities

• Apply state-of-the-art techniques to complex and challenging domains.
• Work with researchers and quants to build flexible ML frameworks for financial applications.
• Optimize training pipelines to utilize HPC resources efficiently.
• Integrate ML models into production systems where latency matters.
• Work across programming languages: C/C++/Python/CUDA and related tools.
• Build scalable ML systems that are observable, performant, and adaptable.
• Other duties as assigned.

Skills

ML research

Python

C++

PyTorch

JAX

TensorFlow

CUDA

HPC

Tools

PyTorch

JAX

TensorFlow

CUDA

Job description

Jump Trading Group seeks research scientists to apply machine learning to state-of-the-art capabilities across complex domains. You will drive end-to-end research from concept to production, iterating on model design, tooling, and infrastructure.

You will collaborate with researchers and quants, optimize training pipelines on HPC resources, and deploy models where latency matters, using Python/C++ and frameworks like PyTorch, JAX, or TensorFlow.