Abhishek de

About Abhishek de

Skilled Applied Scientist with experience in mathematical modeling, machine learning, applied statistics, and experimental design. I am currently working on improving Credit-Risk underwriting for marketplace lending.Programming languages -> Python, SQL, Shell, MATLAB, C/C++ Machine learning -> Regression and classification (GLM, GBM, SVM, Neural Networks), Clustering (DBSCAN, PCA, t-SNE, UMAP, k-means clustering), anomaly detection (LOF, Isolation Forest, OneClassSVM).Theory -> ML, Anomaly detection, Applied statistics, A/B testingTools/Frameworks -> Jupyter lab, VS Code, Git, AWS, Snowflake, Starburst, Docker, JIRA, Confluence, Asana, NotionBig Data management tools -> Apache Spark, H2OOrchestration -> Airflow, Flyte, KubeflowBI -> Looker, Excel

Experience

1. Senior Machine Learning Engineer

   Affirm

   Jan 2025 — Present · Seattle, WA, US

   Applied Machine Learning Engineering team member- Building models to improve 1. credit risk underwriting for point-of-sale originations and 2. return-on-assets based decisioning (cashflow calibration and risk-ranking)- Researching and deploying GBM based models using fine-tuning, transfer learning and stacking- Researching and deploying transformer based foundation models for credit-risk decisioning- Leading loan exploration based reject inference initiatives for expanding Affirm\'s portfolio

Education

• Indian Institute of Technology, Madras

  Bachelor's degree

• University of Washington

  Doctor of Philosophy (Ph.D.), Neuroscience

  2014 — 2020

• Indian Institute of Technology, Madras

  Integrated Bachelors and Masters (Dual Degree Program), Electrical Engineering (B.Tech), Applied Mechanics (M.Tech)

  2009 — 2014