I’m an AI Engineer @ EPAM Systems. I have 8+ years of experience in developing software of various kinds - mostly, machine learning solutions. Currently, I’m developing libraries for efficient LLM inference on Tenstorrent hardware. Previously, I worked on graph optimization and inferential analysis of Locus Robotics warehouses, built AI assistants for the automation of conversational experiences of clients @ Alif 🇺🇿, and did research on machine learning for drug discovery, incl. chemical foundation modeling and multi-task tabular / graph learning, @ Romanovsky Institue of Mathematics. I received my Bachelor’s in Applied Math & CS from Lomonosov Moscow State University in Tashkent.

I’m broadly interested in production-oriented machine learning, esp. in solutions for making predictive systems efficient, reliable, and practically deployable.

I try to be active on social media incl. Medium, Kaggle, and Github by sharing my thoughts and experiences on the field. Check out the references & pages on this website and feel free to reach out by any means that suits you best.

Recent News

• 2026/09/07. 📝 Read about how we develop our vLLM plugin for Tenstorrent hardware in the vLLM blog: Serving LLMs on Tenstorrent Hardware: Inside the vLLM TT Plugin.
• 2026/02/22. 🎉 scikit-fallback v0.2.0.post1 is out! It brings better docs, major features like dynamic ensembling with ThresholdCascadeClassifier and other estimators, and several small improvements.
• 2025/08/13-15. Mentored the runner-up and multiple other teams @ IT Park AI Hackathon!
• 2025/05/01. Started my new journey as a Data Sci / ML Eng @ EPAM!
• 2024/11/08. Finally retook the TOEFL and scored 108, marking a huge improvement from my previous 95 in October 2021!
• 2024/10/26-27. Mentored @ Women Techmakers Tashkent. See my Smooth & Fresh Introduction to Tabular Competitions (w/ scikit-learn, skrub, mlxtend, and feature-engine).
• 2024/10/12. 🎉 scikit-fallback v0.1.1.post0 is out! See also the documentation and the v0.1.0 release w/ bugfixes and new features such as anomaly-based fallback classification.