Machine Learning Engineer Underwriting
Affirm
Build real-time underwriting models with Python and classification expertise at Affirm while working remotely across eligible Canadian provinces. Your work will assess repayment risk. As a Machine Learning Engineer II on the Underwriting ML team, you will develop systems for transaction decisions. You will collaborate with ML engineers, platform partners, and stakeholders to move models from prototypes into production. The team uses tabular and sequential data, proprietary and third-party signals, and strong monitoring as conditions evolve. Key Responsibilities:
Develop and iterate underwriting prediction models Build feature pipelines and training datasets Run offline experiments and productionize selected approaches Integrate models into batch or real-time systems Monitor model and data health Requirements:
Two or more years of machine learning engineering experience Strong Python and production-quality coding skills Classification modeling experience, preferably LightGBM, XGBoost, or CatBoost Experience with PyTorch and distributed processing frameworks Bachelor’s degree or equivalent practical experience Apply machine learning, Python, and monitoring skills to improve Affirm’s real-time underwriting decisions remotely in Canada.
#J-18808-Ljbffr
Develop and iterate underwriting prediction models Build feature pipelines and training datasets Run offline experiments and productionize selected approaches Integrate models into batch or real-time systems Monitor model and data health Requirements:
Two or more years of machine learning engineering experience Strong Python and production-quality coding skills Classification modeling experience, preferably LightGBM, XGBoost, or CatBoost Experience with PyTorch and distributed processing frameworks Bachelor’s degree or equivalent practical experience Apply machine learning, Python, and monitoring skills to improve Affirm’s real-time underwriting decisions remotely in Canada.
#J-18808-Ljbffr
Reference: WJ-3875_16329450 · Posted 15 Nov 2026, 10:00 PM