Machine Learning Researcher

NxT Level • San Francisco, CA

1d • Onsite

About The Position

The Machine Learning Researcher will rethink how credit risk is modeled and evaluated at a fast-growing fintech company transforming credit underwriting across Latin America. This research-heavy role devotes approximately 70% of its time to developing and experimenting with new machine learning architectures and 30% to productionizing successful models. The work aims to uncover predictive signals in complex, real-world financial data and deploy novel models that influence lending decisions at scale.

The researcher will explore approaches beyond traditional XGBoost and gradient-boosted models, including potential LLM and multimodal methods. A background in fintech or lending is not required; experience modeling complex, noisy systems in fields such as robotics, physics, chemistry, or other quantitative disciplines is relevant. The role works closely with the CTO, the Head of Buy Now, Pay Later (BNPL), and a highly technical team of approximately 20 people in San Francisco. The position is onsite five days per week.

Requirements

• 5–10 years of relevant experience in machine learning research, including qualifying academic or PhD research.
• Master's or PhD in Computer Science, Mathematics, Physics, Engineering, or another quantitative discipline from a top-30 university.
• Deep hands-on experience with classical machine learning, statistical modeling, and predictive algorithms.
• Strong Python programming skills for research, data science, and production machine learning.
• Experience modeling complex systems with noisy data, confounding variables, and uncertain outcomes.
• Strong understanding of model evaluation, calibration, temporal validation, and leakage detection.
• Ability to independently design experiments, test hypotheses, and develop new model architectures.
• Experience translating research into production-ready ML systems.
• Willingness to work onsite in San Francisco five days per week.
• Strong foundations in classical machine learning and predictive modeling; this is not an LLM-only research position.

Nice To Haves

• Experience with LLMs, multimodal architectures, or representation learning.
• Research experience in robotics, physics, chemistry, or other complex scientific domains.
• Experience working with large proprietary datasets.
• Experience developing novel modeling approaches beyond established algorithms.
• Experience deploying research models into production environments.
• Exposure to credit risk, underwriting, financial services, or fintech.

Responsibilities

• Research and develop novel machine learning architectures for credit risk prediction.
• Explore LLM and multimodal modeling approaches beyond traditional XGBoost and gradient-boosted models.
• Analyze large, noisy proprietary datasets to uncover predictive signals and features.
• Design and run experiments to evaluate model accuracy, robustness, and predictive performance.
• Apply rigorous evaluation techniques, including calibration, temporal validation, and data leakage detection.
• Build and improve predictive models for complex real-world financial systems.
• Deploy successful research models into production.
• Partner with engineering teams on model infrastructure, monitoring, and data provider integrations.
• Collaborate with technical and business leadership to define the future of AI-powered underwriting.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior