I like turning messy requirements into clean APIs, stable data models, and services that stay fast under pressure
Why it exists: Enable online ordering and delivery for a federal drug store chain through marketplace and delivery partners.
How it works: Backend service integrating internal order system with Yandex Delivery API; handled order status synchronization, delivery updates, and failure retries via RabbitMQ workers. Designed to tolerate inconsistent external callbacks and network errors.
Tech used: Python, Django, PostgreSQL, RabbitMQ, REST APIs.
Why it exists: Provide management and regional offices with consolidated sales, stock, and pharmacy performance reports.
How it works: PostgreSQL-based DWH with multiple large fact tables (400M+ rows). Partitioning applied to manage growth. Complex SQL (CTEs, aggregates, window functions) used for reporting and monthly financial reconciliation.
Tech used: PostgreSQL, SQL, Python.
Why it exists: Synchronize product catalog, pricing, and stock data between legacy Java-based systems and new Python services.
How it works: Gradual migration from Java modules to Python services; API-based communication and async processing via RabbitMQ to avoid blocking critical retail operations.
Tech used: Python, Java (legacy), PostgreSQL, RabbitMQ.
Why it exists: Collect competitor pricing and product availability data for analytics and pricing adjustments.
How it works: Automated parsing tools with scheduled runs; data cleaned and loaded into internal reporting systems. Focused on resilience against layout changes and incomplete data.
Tech used: Python, BeautifulSoup, PostgreSQL.
Why it exists: Maintain stable deployment and operation of backend services across staging and production environments.
How it works: Dockerized services; server configuration and monitoring setup; ensured reliable deployments and minimized downtime during updates.
Tech used: Docker, Linux, Nginx, PostgreSQL.