Machine Learning Researcher

Digital Skills USA; New York (NY); On-site; USD 900,000 - 1,100,000; Full time

Job summary

Digital Skills USA seeks a Machine Learning Researcher with strong mathematical foundations and practical deep learning expertise to join a leading quantitative trading business.

You will develop and test models using complex financial data, combining research judgment with hands-on engineering, supporting trading research across fixed income, currencies and commodities.

Qualifications

• Strong foundations in probability, statistics, algorithms and mathematical reasoning.
• Hands-on experience developing, training and evaluating machine learning or deep learning models.
• Proficiency in Python and PyTorch, with the ability to write clear, maintainable research code.
• Rigorous experimentation with clear explanations of why an approach works.
• Experience working with models in a production environment.

Responsibilities

• Research and develop statistical and deep learning models to identify patterns in complex datasets.
• Apply probability, estimation and regression to evaluate signals and test ideas.
• Build reproducible research workflows in Python and PyTorch.
• Investigate model behavior, diagnose training instability and improve generalization.
• Explore pre-training, fine-tuning, regularization and architectural choices.
• Evaluate trade-offs between model quality, cost and practicality.
• Collaborate with researchers and engineers to deploy research into production.

Skills

Probability & statistics

Deep learning

Python

PyTorch

Research methodology

Production ML experience

Education

PhD in Mathematics, CS or Statistics

Master's with strong research experience

Tools

Python

PyTorch

Job description

Machine Learning Researcher — Deep Learning

Location: New York or Miami

Compensation: Target total compensation of $1m–$3m per annum, with flexibility for exceptional candidates.

Digital Skills USA is recruiting for a leading quantitative trading business seeking a Machine Learning Researcher with strong mathematical foundations and practical deep learning expertise.

You’ll develop and test models using complex financial data, combining research judgement with hands‑on engineering. The work supports trading research in fixed income, currencies and commodities, where noisy datasets and changing conditions create challenging modelling problems.

Previous finance experience is not required. This is an opportunity to apply research developed in technology or AI to a new domain.

What you’ll do:

• Research and develop statistical and deep learning models to identify useful patterns in complex datasets.
• Apply probability, estimation and regression methods to evaluate signals and test research ideas.
• Build reproducible research workflows in Python and PyTorch.
• Investigate model behavior, diagnose training instability and improve generalization.
• Explore pre‑training, fine‑tuning, regularization and architectural choices.
• Evaluate the trade‑offs between model quality, computational cost and practical implementation.
• Collaborate with researchers and engineers to turn successful research into production applications.

What we’re looking for:

• Strong foundations in probability, statistics, algorithms and mathematical reasoning.
• Hands‑on experience developing, training and evaluating machine learning or deep learning models.
• Proficiency in Python and PyTorch, with the ability to write clear, maintainable research code.
• A rigorous approach to experimentation, including testing assumptions and explaining why an approach works.
• Experience working with models in a production environment.
• Typically 3–10 years of relevant experience, with flexibility based on research depth and delivery.
• Ideally a PhD in mathematics, computer science, statistics or another relevant STEM discipline. A master’s degree combined with strong relevant industry experience will also be considered.

Experience that would stand out:

• Research into model architecture, attention mechanisms, mixture‑of‑experts models or long‑context approaches.
• Experience improving training stability or scaling larger models.
• Carefully controlled experiments demonstrating measurable improvements in model quality.
• A practical understanding of GPU memory, compute requirements and performance trade-offs.
• Meaningful contributions to open‑source model implementations or research code.

These are areas of interest rather than a requirement to cover every topic.

Your background:

You could come from an AI lab, an established technology business, a startup conducting substantial research or a financial markets firm.

We’re particularly interested in people who combine mathematical depth with evidence of building and delivering models. Academic research experience is welcome alongside meaningful industry production experience.