• Designing and implementing an Large Language Model system that analyzes and improves the alert system outcomes. Conducting research.• Collaborating with product managers and other co-founders to define project requirements, scope, and success criteria.• Mentoring and managing junior team members, conducting code reviews, and providing technical guidance.

- Developed, deployed, analyzed several machine learning models (predicting targeting dates, players lapse predictions and games recommendation systems). Ensured that models run in production correctly.- Conducted data analysis and visualization. Cleaned, and manipulated raw data. Created graphs, charts, technical reports, detailing data, pre- and post-modelling analysis results.- Collaborated with various teams to search for new project opportunities, and insights and to ensure the model performance and quality of data is high. - Communicated with the key stakeholders regarding progress, timelines, blockers and requirements.

• Designed and developed ML system using LSTM and RL models to augment heuristics-based decision-making system used in high availability software. The task of the ML model was to predict actions based on the current environment state. Drove testing, training and evaluation of ML model.• Conducted feature selection. Optimized data fetching, cleaning, and preprocessing, resulting in a 10x speed increase.• Implemented model evaluation, performance metrics, tunned learning parameters and increased the model’s performance.• Improved decision-making process. Developed custom error injection for LSTM network. As a result, the performance and accuracy ML model significantly increased.• Communicated closely with the client, providing regular updates. Worked on data visualization, technical reports, documentation, and presentations.

Developed a data collection tool that logs to various servers, gets data from different files, writes data into log files, processes log files and outputs data to Kibana. Created Ansible playbooks. Wrote Python program to search through files. Configured several Filebeats. Used Elasticsearch and wrote painless scripts to preprocess log files. Configured Kibana to have proper views. Wrote documentation, helped with organizational tasks and meetings

Technology used: AngularJS, .NET Core, C#, Oracle SQL Developer, Microsoft SQL Managent Studio, Jira, Confluense, ServiceNow, SharePoint, Windows Servers 2008/2012.Helped team in: enhancing Capital Markets Prime Brockerage Customer Trading Platform, finding and fixing bugs, code development, creating weekly and on-demand requests, configuring Windows servers and moving projects from old to new servers, repo migration, writing documentation