Ladan Mahabadi

About Ladan Mahabadi

10+ years as a Machine Learning Scientist demonstrating exceptional success in leading cross-functional teams to develop scalable models that improve efficiency and decision making. Most recently within Amazon Web Services Business Systems, building scalable models for recommendation systems, customer behavior analysis, capacity optimization and generative artificial intelligence. Holder of 9 patents, demonstrating a strong track record of innovative contributions in machine learning and technology. Recognized for ingenuity, cross-functional influence, and a collaborative approach that drives actionable insights and alignment on achieving stakeholder objectives.

Experience

1. Senior Machine Learning Scientist

   Amazon Web Services (AWS)

   Jan 2024 — Present

   AWS BUSINESS SYSTEMS – CAPACITY EFFICIENCY & OPTIMIZATION• Developed and deployed a customized generative AI assistant for AWS Finance, leveraging scalable ML models and NLP techniques. • Designed, implemented, and evaluated novel ML models and prototypes to rank accounts based on critical customer behavior changes, enabling AWS sales and marketing to prioritize outreach and optimize resources. Utilized statistical modeling, unsupervised learning, and quantitative analysis to deliver actionable insights. • Collaborated with cross-functional stakeholders to understand business requirements, define project scope, and ensure successful implementation of ML solutions. • Developed and maintained comprehensive technical documentation, user guides, and project reports, ensuring knowledge transfer and facilitating ongoing strengthening of ML systems. • Automated Bedrock GPU capacity management (Jul ’25): optimization-based system co-built with Bedrock engineering to rebalance GPU capacity across foundation models as demand shifts; cut LSE response times from hours to minutes and automated capacity management for most models; established metrics, dashboards, postmortems, and incident mechanisms.• Reduced throttling and raised utilization via smart GPU capacity loaning and optimal packing.• Supported key product launches worth an estimated $30MM in ’25; projected $6B annual entitlement by 2027

Education

• McGill University

  Doctor of Philosophy Computer Science, Computer Science

  2006 — 2013

• The University of British Columbia

  BSc, Honor's Computer Science

  1999 — 2003

• Simon Fraser University

  MSc, Computer Science

  2003 — 2006