● Implemented platform architecture with the Head of AI. Partnered together with a research team. ● Rewrote and enhanced deployed ML services clusters based on Ray and Nvidia Triton. Developed framework and CI pipelines for CV researchers around ClearML and Kubeflow. Created a model registry and training workflow logging and orchestration. ● As a result, the research team experienced significant improvements in developer efficiency.

● Communicated with Frontend developer, project manager, and CEO to identify and explore possible new AI features for existing mobile applications. ● Reverse-engineered and implemented standalone microservices with docker-compose and K8s, serving as FastAPI endpoints for facial expression, Text2Speech, and Speech2Text. ● The integration of these services resulted in a 12% increase in user app retention.

● Collaborated closely with senior data scientists, the CTO, and the COO to enhance performance and quality of the genetic data pipeline. Contributed to Federated learning research efforts. ● Implemented parallel processing and introduced new workflows in Snakemake for large-scale data analysis. Made system tests through Github Actions. Conducted EDA and PCA analysis on genomic data, utilizing Seaborn and Plotly. In addition, fine-tuned boosting models. ● The pipeline was successfully integrated into another genetic company's workflow, proving its capabilities for handling large-scale data and effectively marketing it to major genetic companies. The Federated learning research culminated in the proof of concept B2B product.

ФББ, Биоинженерия и биоинформатика

Online course Computational Genomics: Hands-on course on Machine Learning for Genomics, Computational Genomics in R