- Created an Analytical Database and set up ETL from the Production Database. As a result of schema upgrades, filtering, and data transformations, data storage space was reduced by 10 times. These changes led to queries being 5 times faster and significantly improved the work efficiency of the entire data science team.- Analysed transactional data and built NLP models for transaction classification. Developed features based on Open Banking Data to improve decision accuracy and lower the number of defaults.- Updated the architecture and retrained the in-house classifier using fresh data, after identifying and fixing issues in the old model. This led to a 4.5x increase in performance speed and a 19% improvement in accuracy.- Identified problems with external API response times for requests with Open Banking data. Solving those resulted in 3 times fewer timeouts and 50% API response time reduction.- Led the integration of Open Banking with two of the largest UK loan aggregators. As a result, customers using Open Banking are up to 10 times more likely to convert to disbursal.- Oversaw several integrations with major Open Banking providers. Identified issues and bottlenecks that arose. Communicated the problems to the providers and our development team to resolve backend issues. Fixed issues related to classification.

- Developed a novel deep learning architecture for semantic segmentation using PyTorch. The model (Inceptioned U-Net) for embryo component detection with a mean IoU score of 86.56% on validation data and 70.16% on unseen data.- Authored a research paper accepted to Fertility 2023, a major British fertility conference.- Collaborated with a team of 3 Machine Learning researchers in the development of the model.

- Developed an in-house recommendation model with high accuracy and speed. The model replaced a third-party solution, saving the company thousands of dollars annually and improving the user experience.- Achieved a 9.6% increase in prediction accuracy, leading to more relevant recommendations and higher user engagement.- Software developed by me calculated A/B test statistics for 1.1 million users monthly, enabling data-driven optimisation of recommendation algorithms.- Implemented logic for automatic calculation of A/B test metrics, enabling data-driven decision-making for recommendation model improvements.- Developed tools for analysts to create user segments for tests, enhancing flexibility and safety in testing processes.- Implemented monitoring solutions using Kafka, ensuring smooth system operation and real-time performance evaluation.

- Reduced sprint planning time by 15% by developing an NLP model that matched employees with a task in Jira.- Saved 10% of resources and time by developing REST API micro-services using Python to automate routine analytical tasks.- Worked with big data (tables with more than 150M rows).- Wrote efficient scripts in SQL and Python to extract insights from large datasets with sensitive financial data in ClickHouse and PostgreSQL database systems for Marketing and Business departments.- Introduced a data-driven culture in the company: created documentation on how to use data for different roles in the company and educated 6 people from Marketing and Business departments.- Deployed, maintained and monitored production code on Linux servers using Docker and BitBucket.

- Worked directly with clients to understand their unique challenges and find creative solutions to address them.- Used analytical systems from Google for web analytics: Google Analytics and Firebase to monitor web and mobile statistics, and Google Tag Manager to create and track custom events.- Created dashboards in Power BI for McDonald’s Eastern Europe.- Created a custom attribution model for a client: used Google Big Query to collect data and Python to analyse it. Helped reduce the client’s marketing budget by 5%.

- Led the growth of a product from inception, to 80,000+ registered users and 1,200,000 monthly peak site visits.- Assisted 3 YouTube streamers in achieving 100K subscribers within 2-3 months.- Built and managed a cross-functional team of 6 professionals, advocating for gender equality by hiring female software developers.- Managed 6 people (Graphic Designer, Team Lead, 2 Software Engineers, SMM and Accountant).- Oversaw all facets of operations – finances, legal, marketing, outreach, etc.

- Worked in a team to identify exciting FinTech start-ups and organised company presentations to UCL students.- Developed skills in professional communication, event management, and organisation by arranging a cryptocurrency panel attended by 3 start-ups and 200 attendees.

- Worked face-to-face with fresher students to help them adapt to university life.- Developed important organisational and communication skills, when scheduling weekly events with a group of students in my spare time from university.

- Learnt about Transition Metal Dichalcogenides and other graphene-like 2D nanostructures.- Performed scientific research and produced an overview report on the uses of Transition Metal Dichalcogenides.