Arina Shkanova Email & Phone Number

Berlin, De

Ozon is one of the first e-commerce companies in Russia with more than 9 million products across more than 20 product categories including clothing, groceries, home goods, and electronics. The volume of orders of more than 32.2 million in 2022.Key Projects:1) Project for Client Service Analytics Team (managing project for evaluating the quality of client services (defect rate), developing dashboards, creating new ways for ).- Created new ways for increasing accuracy in determining responsibility for defects in working collaboration with business analyst;- Supported an algorithm with more than 5-6 sources and 10 steps ;- Created a pipelines with Airflow for everyday updating;- Developed an interactive dashboard in Superset containing graphics, and tables with all information for managers to process information about defects;- Optimized the code of the algorithm to work faster in Vertica( from processing the information (400 million rows) and calculating metrics for 45 min to doing it for 15 min);2) Pet project:“E-commerce” analytic project for online store including analysis of statistical values within a time series, сohort analysis of users, and RFM user segmentation;

Analytic project for Fuel Systems Team.(Creating algorithm for descent duration statistic reports)- took a part in developing algorithm for determining start point of descent on Python (Pandas, Numpy, Matplotlib);- use additional data from TeraData source;- created algorithm for cumulative graphs;- developed algorithm for creating reports in excel with python lab xlsxwriter;- took a part in developing algorithm for determining start point of descent on SparkAnalytic project for managers of manufacturing engineers team.-took part in the creation of a program written in VBA (excel), which helped managers keep track of the work of team members.

Group projects: 1. "Smart Scales" is a service for recognizing goods that do not have barcodes. Detection of goods (fruits / vegetables), determination of the type, weight and total cost of goods. - I was engaged in the selection of the optimal model for product detection. - Collected and processed data for training YOLOv4 in additional classes. - Worked with the darknet framework. - I selected and applied an algorithm for recognizing numbers on the scales. - Created a web application in flask. https://github.com/arinashkanova 2. "Python helper" - a telegram bot that knows the answers to the most common questions about python. - Collected a database of answers and questions. - I found the best way to encode information and a way to find the best answer based on the similarity to the question. - Created a telegram bot. https://github.com/arinashkanova 3. Recognition of the category of the object of the museum fund by description or image. All Cups (Museum Fund) (Hackathon) - Worked with text data. Used nlp for classification. - Applied the staking method for classification with a textual description of the exhibit and photos. Personal projects: - Participated in the kaggle.com competition (Predicting the cost of houses by features )

Responsibilities:- Creating the technical processes. Control of delivery.- Formation of the base of technical sheets.- A quick fix on the troubles in manufacturing.- Design of those refurbishment processes.Achievements:My participation in the working team projects has made it possible to achieve the best quality in all metrics.During my second year, I oversaw a new project (Insulation blankets) on my own - directly with the American side.