Reem Mahmoud is the Director of Data Science at intervu.ai, where she leads the development of machine learning systems that power the company’s products. Her specialty is production ML search, and she has worked deeply on the challenge of moving search from prototypes to systems that scale and stay relevant.

Her expertise spans the modern search stack: vector search fundamentals such as embeddings and embedding pipelines, hybrid architectures that combine vector similarity with filters, recency, and business constraints, and scalable indexing built on concepts like inverted indexes and Lucene. She also works with multimodal embeddings across text and images, feature fusion for richer relevance, and the practical trade-offs between LLMs and specialized encoders, along with the vector database decisions that follow.

As a leader, Reem focuses as much on process as on models: offline testing, A/B metrics, and enabling engineers to iterate quickly so that search improvements balance latency, relevance, and business KPIs. She shares this experience through talks and podcast conversations on production search, embeddings, and machine learning.