Looking for a job

Ksusha Malkova

@ksumarshmmallow

Message

Data Scientist specializing in recommender systems, retrieval, and ranking—boosting GMV via A/B-tested models.

Russian Federation

Message

What I'm looking for

I want to work on production recommender systems where I can improve candidate generation, cold-start, and ranking/retrieval using offline evaluation and automated A/B testing to deliver measurable business gains.

I’m a Data Scientist working on recommender systems, ranking, and retrieval at Wildberries, and I genuinely love this field. What excites me most is the idea that recommendations can shape how people discover things — not just by making systems more accurate, but by making them feel more alive, more personal, and more engaging for users. I’m especially drawn to new approaches that are only becoming possible now through new architectures and advances in user behavior modeling.

I truly enjoy exploring new ideas, discussing them with others, and turning them into real product decisions. I also love sharing what I learn, whether through teaching, educational initiatives, or simply conversations with people who are passionate about similar topics. Recommender systems are special to me because they are much more than just a technical domain: they combine technical depth, creativity, and direct user impact — and that is exactly the kind of work that excites me most.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Wildberries

Jun 2024 - Present (1 year 10 months)

I was a data scientist in Wildberries, recommendation and search platforms, developing neural recommendations and ranking systems

Python PyTorch SQL Spark Hadoop Linux Airflow Apache Airflow Faiss Triton Inference Server Recommendation Systems Ranking Systems

Current

Data Scientist (RecSys)

Current

Wildberries

Jan 2024 - Present (2 years 3 months)

Built and deployed neural recommender system improvements that increased GMV by 13%, recommendation diversity by 15.7%, and ATPU by 7.2% in production A/B tests. Led work across candidate generation, cold-start recommendations, offline evaluation, and training pipeline optimization to improve diversity, relevance, and ranking quality.

Recommender Systems Ranking A B Testing PyTorch Faiss