Contributed to a legacy backend microservices project aimed at providing high-throughput message pipelining, from receiving raw events to delivering formatted messages to users.My primary focus was on:- Addressing performance issues and instabilities within the legacy backend microservice architecture, particularly in the high-demand message pipeline.- Eliminating unnecessary operations and optimizing the way we save and update information in the database.- Conducting extensive refactoring of these services, enhancing their maintainability and facilitating the development of new features.- Implementing key performance metrics for improved visibility and understanding of system operations without overloading the infrastructure.- On-call duties on a weekly rotation basis (1 week on/1 week off), ensuring system reliability and quick response to issues.- Significantly reducing the frequency of critical issues in managed services.Additionally, I implemented numerous minor and major improvements, responding effectively to a wide range of feature requests.Technical stack: Django, DRF, Redis, RabbitMQ, PostgreSQL.

Engaged in the development and support of various web services for an asset management hedge fund, including the creation of a lightweight alternative to a complex legacy system. This new system efficiently stores various types of metadata for asset management and facilitates authorization and policy management with a m2m (many-to-many) parenting system and flexible querying capabilities.Additionally, I conducted extensive refactoring of several legacy systems, focusing primarily on improving the code base. This effort was aimed at making the code more manageable and reducing the difficulties associated with working on these systems.Technical stack: FastAPI, kafka, Peewee, asyncio, PostgreSQL.

Developed various internal backend products from scratch, including:- A service for antivirus signature database development, significantly easing and accelerating the process of developing and maintaining antivirus signatures.- A service for antivirus testing across a large file set (approx. 10 million files).- A semi-automatic service for processing incoming antivirus events (false positives and false negatives triggering, etc.), greatly simplifying the analysis of excessive incoming event streams, previously unmanageable manually.- A service for the development of a rules database for the company's security product.- A vulnerabilities alert service.- An internal library comprising diverse tools to facilitate the development of similar services.Technical stack: Tornado, Peewee, asyncio, PostgreSQL, ClickHouse, redis.