Leonid Murzinov Email & Phone Number

Project: Migration from NFS and ClickHouse to Snowflake• Rewrote Spark pipeline components in Scala for Snowflake integration.• Developed AQA tests in Java.Tech stack: Snowflake, ClickHouse, NFS, Airflow, Python, Scala, SparkProject: Pipeline Development for Google BigQuery (GBQ) service• Enhanced pipeline components in Python for optimized data ingestion and processing with (GBQ).• Redesigned and optimized Airflow DAGs to ensure efficient orchestration of the entire data pipeline.Tech stack: Google BigQuery, Airflow, Python

Development and implementation of pipelines based on a new data storage system for testing and training models:• Developed and implemented a complete Airflow pipeline for data processing: from raw video files to annotated images.• Created tools for uploading, downloading, processing, and initial annotation of data using neural networks.• Developed a system for automating data collection and annotation, accelerating the processing workflow.• Contributed to the creation of a data visualization system, aiding in accuracy improvement for training and error identification.Tech stack: Python, Airflow, Spark, Docker, Luigi, FFmpeg, Superset, Streamlit, DBTDevelopment of API and SDK for data management in model training and testing processes:• Expanded the functionality of API and SDK, simplifying data and annotation handling for data scientists.• Provided support and development for existing systems, ensuring their stability and efficiency.Tech stack: Python, PostgreSQL, FastAPI, CephLeadership in transitioning to a new data storage and processing system:• Actively participated in the design and implementation of a new data storage and annotation system.• Managed the project for migrating data to the new system, ensuring efficiency and minimizing risks.• Developed and implemented tools for data migration, ensuring a safe and smooth transition.• Conducted training demonstrations for the team on working with the new system.

Developed a data processing service from scratch, including extracting, filtering, and aggregating data from multiple data warehouses (DWHs) such as PostgreSQL and MS SQL, using advanced SQL scripts.• Created a specialized module for detecting errors and discrepancies in data, utilizing Python with pandas and numpy libraries.• Developed and implemented automated data verification processes using Apache Airflow, ensuring regular data updates and quality control.• Facilitated the integration of the entire application into Docker containers, significantly simplifying the testing and deployment processes on servers.• Developed and deployed a Flask-based web application for data management, including capabilities for user data correction, error note creation, ticket submission, and notification distribution.Tech Stack: Python, Airflow, Docker, Flask, HTML, CSS.Led the organization of ETL processes in a project aimed at modeling the optimal distribution of equipment among agricultural enterprises.• Handled data of various natures (including relational and non-relational databases, images), performing extraction, filtering, and aggregation.• Developed a module for cleaning and transforming data for an optimizer solving integer linear programming problems.• Created a database and a module for uploading data in the required format.Tech Stack: Python, SQL, Pyomo.

Москва, Россия

Actively participated in a project for the detection and recognition of vehicle license plates and types.• Conducted data scraping using Beautiful Soup, Selenium, and APIs.• Performed image annotation and preprocessing for neural networks, using various methods and libraries.• Deployed and tested neural networks on servers using Docker.• Optimized neural network performance metrics.Tech Stack: Python, YOLOv4, Pillow, Pandas, Selenium, Beautiful Soup, Docker.Developed a service for face detection and recognition.• Developed infrastructure for working with the Insightface library, including API integration and web interface handling.• Created a Flask-based web application for uploading and processing images.• Determined and improved face recognition quality metrics.• Prepared and processed datasets for neural network training.

Магистр

Education record