Duties:* Conduct data preprocessing, feature engineering, and EDA on large-scale datasets* Conducted feature engineering to extract meaningful features for the models* Build, evaluate, and submit for deploy ML models* Collaborate with the team members to adjust the input / output to fit the joint infrastructureAchieved:* Successfully developed and deployed ML models for CTR and lead rate prediction* Stabilized and outperformed the baseline (by 20-30%) CTR during the bidding process* Developed the pipeline to improve MLOps lifecycle proceduresTools:* Development: python(numpy, pandas, scikit-learn), modelling (catboost, xgboost, deepctr, FFM), deep learnig (torch)* Deployment: mlflow, airflow, docker* DB & Cloud: aws, clickhouse, mysql

Develop optimization and ML models to get the business insights. Manage and direct the team to ensure efficient task distribution and timely delivery of results to upper management.Duties:- Build business optimization models with the objective to reduce costs / increase incoming cashflow by getting optimal production volumes, modes and delivery routes.- Develop ML models to submit the results to the optmization blockAchieved:- Within the optimization block developed and delivered the insights to the management and BoD consideration:-- the optimization model for Kuzbass-Khakasia mines optimal production and distribution;-- lignite coal open-pits optimal production and distribution (subject to the consumers demand and logistics restrictions);-- lignite coal optimal production and distribution with the multiple demand scenarios.- Within the ML block developed the washing facilities yields prediction model based on the incoming product characteristics and the equipment.Tools:- Optimization: gekko, cvxopt, pandas- ML: tensorflow.keras, pandas

Contribution to the methodology and the toolset of a powerplant's life-cycle economics calculation.Duties:- Equipment price prediction based on its technical characteristics.- Research of the specific equipment dependency on the market indices.- Development of the economics methodology for the software development based on the cost functional analysis.Achieved:- The model used as the test case to reduce the time required to get economics while designing the power-plant elements.- Cost functional analysis methoodology deployed for construction a power plant.Tools:- Development: tensorflow.keras, xgboost, catboost, scikit, pandas, autogluon- Deployment: flask

Contribution to the business development activities (oil & gas production).Duties:- Development of the financial and strategic KPI-s models for the subsidiaries.Achieved:- The models and the derived strategy recommendations were delivered for the management decisions.