At Meta, I work on Ads Recommendation & Ranking initiatives to improve ad relevance, personalization, and ranking quality. I use deep learning, embeddings, and behavioral signals to strengthen user-ad matching and candidate quality.
I also shape experimentation and causal measurement across ads initiatives, using A/B testing, power analysis, and treatment-effect estimation to inform product and monetization decisions.
At Meta, I’ve applied LLMs, RAG, semantic search, and vector retrieval to advertising workflows, and developed evaluation frameworks for retrieval relevance, groundedness, safety, and model drift.
Previously, at AllState, I developed customer loss and retention models and created a production scoring pipeline with an expected annual savings impact of approximately $1.5 million. I also contributed to a modeling framework for actuaries working with distributed data.
