• Contributed to the development of an AI-based Evidence Analysis tool that extracts, identifies and classifies objects from images.• Engineered and fine-tuned a ResNet-based model achieving a 93% accuracy rate in detecting alcohol presence in images (Python, TensorFlow, NP).• Developed product’s backend code to establish model pipelines and facilitate output generation, enhancing operational efficiency.• Delivered a presentation on project methodologies and outcomes at the BU Experiential Learning Expo.

• Empirical analysis and creation of investment memos for startup companies• Development of data analysis scripts for startup memos review (Python)

• Design of neural networks for long-term crypto investment strategies (Python, TensorFlow)• Development and analysis of empirical methods of finding correlations between crypto prices and fundamental parameters• Collaboration with company’s financial analytics and traders when designing input parameters and algorithms

• Developed and tuned a neural network (366 neurons, 52991 parameters) that uses economic indices from PWT data to assess the stability of political situation in a country with an accuracy of 70.48% (Python, NumPy, TensorFlow); research paper will be posted on Bloomberg.com• Developed a code for parsing Twitter profiles based on the keywords of their descriptions using Twitter’s API (Python)• Implemented data extraction and cleaning scripts for database creation (Python)• Was the head of intern developers’ group, pitched the project to the company’s head management

Intern at QIWI plc. Worked as a content producer, was working on promoting the QIWI FINTEEN department in social networks, as well as creating contests for company's promotional events.