AI Researcher: Financial Time Series & DL Architect

Goldman Sachs; New York (NY); Presencial; USD 100 000 - 200 000; Tempo integral

Resumo da oferta

Goldman Sachs AI Research – Associate in New York, NY seeks a deeply hands-on researcher to design, train, and evaluate deep learning models for financial time series. Lead end-to-end research from data curation to production, with multi-node GPU training, rigorous evaluation, and collaboration across quant and engineering teams.

You will own model architecture, training experiments, and evaluation methodologies, while mentoring juniors and contributing to a firmwide research platform.

Qualificações

• Bachelor, Master, or Ph.D. in Computer Science, ML, Statistics, Mathematics, Physics or related quantitative field.
• 3+ years industry experience building, training and deploying deep learning models, with time-series focus.
• Expertise across CNNs/TCNs, Transformers, autoencoders, GANs, diffusion models, GNNs, RL.
• Strong knowledge of ARIMA, GARCH, Kalman, state-space models.
• Proficiency in PyTorch, TensorFlow/Keras and/or JAX/Flax.
• Experience with distributed/multi-node GPU training and ONNX deployment.
• Excellent communication and adaptability in evolving environments.

Responsabilidades

• Design, train, and evaluate deep learning models for financial time series end-to-end.
• Develop scalable training loops across multi-node GPU clusters and cloud environments.
• Build robust evaluation frameworks with walk-forward validation and regime analysis.
• Collaborate with quantitative researchers and engineers to productionize models.
• Mentor junior researchers and contribute to a shared research platform.

Conhecimentos

Deep learning

Time series

Distributed training

PyTorch

TensorFlow

JAX

Formação académica

Bachelor/Master/PhD in CS/ML/Quant Finance

Ferramentas

ONNX

Docker

Kubernetes

Descrição da oferta de emprego

Goldman Sachs AI Research – Associate in New York, NY seeks a deeply hands-on researcher to design, train, and evaluate deep learning models for financial time series. Lead end-to-end research from data curation to production, with multi-node GPU training, rigorous evaluation, and collaboration across quant and engineering teams.

You will own model architecture, training experiments, and evaluation methodologies, while mentoring juniors and contributing to a firmwide research platform.