Ксюша Малкова

I’m a Data Scientist specializing in recommender systems, retrieval, and ranking. At Wildberries, I build and deploy neural recommendation improvements that translate into measurable business and engagement gains in production A/B testing.

I’ve focused on candidate generation and cold-start, including mitigating recency bias in nearline recommendations. By rebuilding the training dataset, switching to a next-item objective, and introducing a multi-target anti-bias loss for candidate generation—plus removing RAM bottlenecks via an iterable dataset pipeline—I delivered GMV +13% and diversity +1.7%.

I improved recommendation diversity by replacing quota-based post-processing with category-aware retrieval. Using multiple FAISS indices split by category and tuning the diversity–relevance trade-off through five A/B iterations, I achieved +15.7% diversity and +1.5–5% on key engagement metrics; I also strengthened iteration velocity by automating h2h A/B experiments into a reproducible workflow.

I continuously evaluate and refine ranking quality—sourcing and integrating item-level features into the ARGUS ranking model and running systematic ablation studies to guide architecture iterations. For new users, I built a cold-start media recommendation baseline using clustered SigLIP embeddings, resulting in ATPU +7.2% in A/B testing and production deployment.