Development of chatbot system for car sales and customer support. Minor Machine learning tasks as server-side PII detection, client-side document detection with memory constrains;Achievements:Scalable and PII-secure chatbot for car sales, that leverage information provided by business;New-Generation Chatbot continuous quality measurement using LLMs and expert data;Implemented document recognition neural network that weights 130 KB;

Project: Building plans recognitionRole:ML Engineer/MLOpsBuilding ML pipeline in an RnD project for generating 3D models of buildings. Transitioning the project from an RnD stage to a Production stage.Tasks: - Designing and implementing a machine learning pipeline in computer vision; - Training team to follow best practices of machine learning and ML models analysis;- Consolidating a large park of ML models and data processing stages into a unified structure with learning artifacts and logs management.Technologies: Python, FastAPI, ClearML, Airflow, Kubernetes, OpenCV, Gitlab CI, GCP.MLOps in CV (Building Plans Recognition)Project: Development of an Intelligent Plant Growth Advisor (based on images and text)Role: ML EngineerTasks:- Development of models for the classification of plant diseases;- Development of models for extracting named entities from text corpora about plants.Achievement: - Developed a PoC proving that recognition of certain plant diseases is possible at an acceptable level.Technologies: PyTorch, Fast.ai, ClearML, PostgreSQL, NLP, NER, CV.Project: Uniswap Analytics on GCPRole: Backend developerWriting an analytics service for evaluating the balance of cryptocurrency wallets on Uniswap.Tasks:- Designing application architecture;- BigQuery database management;- Development of dashboards in Google Looker Data Studio.- Development, deployment, and support of services working with the BitQuery API (cryptocurrency transaction analysis);Technologies: Python, GraphQL, BitQuery, BigQuery, SQL, GCP.

Working on projects with different data types: CV, NLP, recommendations. Developing neural networks for text and visual product catalog search. Building a recommendation system, demand forecasting, optimizing the set of promo actions. T͏e͏a͏m͏: 2 Data Scientists, ML Engineer, MLOps, Team Leader.Achievements:- Increased text search conversion by using neural network vector search, increasing search CTR by 5%;Improved the recommendation system, increasing the CTR on recommendations by 10%;- Achieved an increase in sales in all projects through the implementation of ML solutions or the modernization of existing approaches to data collection and processing;- Transitioned the demand forecasting and promotion optimization system to AWS, resulting in an increase in the development speed, experiment reproducibility and transparency.Building a system of recommendations, demand forecasting, optimizing a sets of promotions. Tasks:- Developing, deploying, and validating models;- Developing production machine learning pipelines;- Data collection and preparation;- Writing tests for various stages of data preprocessing;- Maintaining, refactoring, and enhancing existing codebase; - Project planning, complexity estimation.Technologies: Python, FastAPI, PyTorch, AWS, GCP, GKE, BigQuery, Redshift, XGBoost, CatBoost, LightGBM, NLP, OpenCV, PostgreSQL.

As a data scientist mainly engaged in the analysis of texts to identify problems and points of growth.Team: 3 Data Scientists, 1 Data Engineer, Team Leader.Achievements:- Helped the client to understand their media coverage and receive feedback from the bank;- Successfully used text clustering pipelines in various bank departments for an extended period of time.Tasks:- Summarizing media coverage of the bank (analyzing text from online media);- Text analysis and web scraping;- Analyzing medium-sized text data to find meaningful insights about the bank's operations;- Developing a review clustering pipeline;- Validating scoring models;- Forecasting customer complaints and employee illnesses.

Support, debugging, and extending the functionality of desktop and web applications for a medical company.Team: 3 Backend developers.Achievements:- Created reliable automated UI test suite for the common and critical parts of application;- Increased the speed of the ORM and optimized key SQL queries to the database;Tasks:- Writing automated tests for interfaces;- Debugging, fixing, and extending functionality of desktop and web applications for a medical company;- Writing back-end for web services;- Writing and profiling SQL queries;- Refactoring custom ORM in the project.