Scott Oddo

• Designed and led the evals and quality checks for an LLM-powered clinical reporting module within a regulated, FDA-cleared software environment
• Built and deployed a real-time predictive risk-scoring model predicting hypoglycemia onset up to 24 hours in advance, with lower false alarms than any published model
predicting hypoglycemia onset up to 24 hours in advance for hospitalized patients.

Consulted on building an AI-powered personalized shopping assistant using implicit and explicit data.

• Partnered with MIT researchers to develop neural network architectures demonstrating true continuous learning in real-world, non-stationary environments — applied leading-edge ML research toward production-relevant problems

• Built decision tree, random forest, and XGBoost classifiers predicting feature usage during enrollment to personalize onboarding and improve retention
• Quantified feature-level causal impact on first-renewal rates to guide product prioritization and roadmap investment
• Developed customer lifetime value models and segmentation to inform go-to-market and targeting strategy

• Led and mentored a 7-person cross-functional team of data analysts, data scientists, and data engineers
• Built an AI churn-prediction model that reduced churn by 10%
• Developed segmentation, CLV, survival-curve, and price-elasticity models to guide pricing strategy
• Led consolidation of analytics infrastructure and ETL pipelines from multiple data platforms onto a single Oracle-based stack

Built customer segmentation, product recommendation, and behavior-analysis models that improved usability and adoption of the website-builder product. Resolved enterprise-wide data fragmentation that had blocked cross-brand analysis.

Built an AI-driven product recommendation system based on purchase history. Used clustering to identify five key customer segments and correlated customer preference signals to power new personalized recommendation types.

Built a model predicting optimal send-time for client email marketing campaigns. Developed an NLP algorithm identifying the most effective subject-line keywords and generated cross-sell leads that increased add-on sales. Built ETL processes using Hadoop/Hive to integrate and analyze marketing data.

Led global product recommendation and personalization services deployed to 100MM+ devices across 15+ countries and 20+ languages. Grew the Nokia App Store from 1M to 11M downloads/day in 18 months and designed a tool enabling configuration of 20,000+ localized store versions at scale.

Built a scalable, AI-powered media-discovery network delivering personalization and product recommendations across music, movies, video, and blogs. Led a 5-person data science/engineering team plus cross-functional project teams.

Bachelor of Arts in Mathematics from Tufts University.

Grade: 3.76

PhD in Cognitive and Neural Systems from Boston University with a GPA of 3.76.

Master of Science in Mathematics from the University of New Hampshire.