J
John Foley
About
Detail
United States
Résumé
Jobs
verified_user
0% verified
- ASenior Data ScientistAstroSirensJan 2022 - Current (4 years 10 months)• Led end-to-end delivery of production ML systems in Python (FastAPI services + batch pipelines), from problem formulation
and data exploration through deployment, monitoring, and retraining workflows.
• Architected scalable inference and feature pipelines using Kafka-driven event streams and distributed processing patterns,
improving data freshness from daily to near-real-time and reducing time-to-signal by ~70%.
• Built and deployed PyTorch and TensorFlow models across classification/regression use cases with rigorous evaluation
(precision/recall, AUC, calibration) and business-aligned KPIs, increasing target metric performance by 10-20% depending on
product area.
• Designed experiment frameworks and st 
Machine Learning EngineerPinterestNov 2019 - Jun 2021 (1 year 8 months)• Shipped production ML services in Python supporting large-scale recommendation and content understanding workflows,
partnering with product to translate ambiguous goals into measurable online metrics.
• Built feature engineering and training pipelines with distributed data processing and robust data quality checks, improving
training stability and reducing pipeline failures by ~30%.
• Developed and evaluated ranking/classification models using TensorFlow/PyTorch, applying hyperparameter optimization and
metric-driven iteration (AUC, precision/recall) to improve model quality and reliability.
• Implemented real-time and near-real-time scoring patterns integrating streaming signals (including Kafka-style event- Data Science InternCensusSep 2018 - Sep 2019 (1 year 1 month)• Performed data mining and statistical analysis in Python and SQL on large structured datasets to identify anomalies and drivers
of key outcomes, delivering stakeholder-ready insights.
• Built baseline clustering and classification models with clear evaluation (precision/recall, ROC-AUC) and documented
assumptions to support decision-making and future iteration.
• Prototyped reproducible ETL pipelines and data validation checks, improving dataset consistency and reducing manual analysis
effort.
• Presented findings to technical and non-technical audiences, translating model results into actionable recommendations and
next-step experiments.
Education
verified_user
0% verified
- PHD, Chemical EngineeringUniversity of VirginiaSep 2013 - Aug 2018 (5 years)Charlottesville, VA
- Bachelor's Degree, Chemical EngineeringOregon State UniversityJan 2008 - Dec 2012 (5 years)Corvallis, OR