Drove the development of user profiling algorithms, experimenting with advanced clustering techniques and zero-shot classification using Facebook’s BART to map over traveler’s clustered memories to Expedia’s taxonomy, improving personalization.

Optimized the LTM (Long Term Memory) model with Redis-based data storage, enhancing user profile refinement and ensuring seamless integration with the core taxonomy, reducing average system latency by 35%.

Designed and deployed a robust ensemble-based model on Flyte and Databricks, parameterized for scalability, validated with a Pytest suite and cluster silhouette scores of 0.6–0.7, ensuring high model reliability and performance.

Collaborated with Senior Director of Machine Learning Science to align ML models with business objectives, stakeholder requirements, and measurable outcomes.

Drove the development of user profiling algorithms, experimenting with advanced clustering techniques and zero-shot classification using Facebook’s BART to map over traveler’s clustered memories to Expedia’s taxonomy, improving personalization.

Optimized the LTM (Long Term Memory) model with Redis-based data storage, enhancing user profile refinement and ensuring seamless integration with the core taxonomy, reducing average system latency by 35%.

Designed and deployed a robust ensemble-based model on Flyte and Databricks, parameterized for scalability, validated with a Pytest suite and cluster silhouette scores of 0.6–0.7, ensuring high model reliability and performance.

Collaborated with Senior Director of Machine Learning Science to align ML models with business objectives, stakeholder requirements, and measurable outcomes.