Director of Product Management
FULLER Equip · Pasadena, CA (Remote)
July 2021 — Present
- Designed and led development of a Learner Intelligence Store (LIS), an AI intelligence layer that aggregates learner telemetry, behavioral signals, and educational metadata to power hyper-personalized learning at global scale.
- Increased course completion 41% for learners on AI-generated pathways versus the static catalog.
- Implemented vector embedding pipelines and retrieval (RAG) architectures, improving retrieval precision 0.59 → 0.86 (human-rated relevance) through embedding-strategy and chunking optimization.
- Built an eval harness scoring recommendation relevance, raising model pass-rate 59% → 89% across 7 cohorts.
- Reduced inference cost 50% by routing low-stakes recommendations to smaller models.
- Integrated Azure AI and OpenAI models to generate personalized course recommendations and adaptive learning pathways from behavioral signals and learner goals.
- Architected and launched Fuller's Learning Management System (LMS) and Learning Record Store (LRS), modernizing the institution's digital learning infrastructure for non-degree learners worldwide.
- Scaled active learners +81% in 13 months and mobile session engagement +48% through AI-driven personalization and a mobile-first redesign.