● Developed a high-performance desktop app with PyQt6 and Pydantic, implementing complex business logic and data validation. ● Reduced key operational time by 2-3x for the client through process automation and an optimized user interface.

● Enhanced a microservices-based platform by developing new REST API endpoints using FastAPI and integrating Keycloak for secure authentication. ● Optimized database performance by 30-50% by refining complex MongoDB queries and data models. ● Integrated Elasticsearch to handle the most problematic queries, accelerating search and update operations by 2-20x and ensuring data consistency.

● Developed and maintained a core Django/DRF platform unifying workflows for 5+ departments. ● Implemented async task handling (Celery), cutting API response times by 40%+. ● Integrated OpenAI and other APIs to automate key business processes and introduce AI-powered features. ● Designed and launched an AWS-based Telegram bot (3k+ users). ● Configured AWS services (Lambda, API Gateway, DynamoDB, S3, CI/CD, IAM).

AWS-based Telegram bot to notify subscribers about new job openings with the ability to filter received ads by selected technologies and UI for recruiting staff. ● Configured all necessary Amazone web services: API-Gateway, Lambda, AWS Amplify, S3, DynamoDB, CodeCommit, IAM, CloudWatch, CI/CD and others. ● Wrote code that provides the required functionality including interaction with UI written in Angular, in accordance with OOP principles, tested, launched and maintained the service. ● An AWS-powered Telegram bot that notifies subscribers about new job openings. Users can select their preferred technologies during registration to receive only relevant ads. Recruiters can post and manage vacancies through a dedicated web interface. Over 3k subscribers as of January 2024. https://t.me/Azati_bot

● Developed end-to-end Computer Vision solutions, delivering a 3x performance boost for a data processing pipeline and enabling real-time sports analytics. ● Built and productionized ML models (YOLO) for object detection and classification, managing the full cycle from data labeling to inference scripting. ● Solved diverse business problems: automating data extraction from images and creating a real-time penalty detection system for handball.

● Architected and launched an AWS serverless platform that automated document workflow for an accounting firm, boosting team productivity by 30%+. ● Designed the DynamoDB schema and integrated multiple APIs (Telegram, Google Drive) for automated document collection, sorting, and tracking. ● Owned the full development cycle: from infrastructure design (Lambda, API Gateway, EventBridge) to development, testing, and maintenance. ● Developed a Telegram bot designed to streamline the transfer of client documents to the accounting firm's employees. The bot offers a wide range of features, including the ability to register and use multiple companies under a single Telegram account, add employees to a client's company, view statistics, and more. For the accounting firm's staff, the bot provides a separate set of functionalities related to client management and other administrative tasks.

● Developed a 24/7 Telegram bot for a real estate agency that acts as a virtual assistant. The bot converses with users to answer questions about services and find suitable properties by clarifying their needs in a natural dialogue. ● Engineered a scalable AI assistant for real estate using RAG architecture with OpenAI API and Milvus vector database. ● Built the backend with FastAPI and containerized the entire system with Docker, enabling 24/7 client support and property matching. ● Delivered a production-ready MVP with customizable prompts and semantic search logic.

● Built a desktop trading platform with PyQt5 GUI, featuring real-time market data processing and order execution. ● Developed core trading engine with risk management, portfolio tracking, and multi-account support.