Anna Kuznetsova Email and Phone Number

 Workout of customer profile & decision criteria in Moscow Region townhouse purchase process on the basis of Shap Values CRM system analysis.  Conversion of business tasks into algorithmic ones. Practical implementation and support of machine learning models. Training company employees to understand ML algorithms.Expected business effect:● Workout of highly informative and effective management accounting.● Construction, advertising, financial and marketing policies efficiency improvement by identifying the profile features that are most important in the decision-making process by the client.

 Bank daily transactions control on the basis of ETL process Fraudulent & questionable transactions identification & control from relational DBMS data (MySQL)Business effect:• Standard approach to banks transactions analysis and control risk was worked out. • 700 transactions with materiality level of 3.5 mln RUB were identified. It potentially could lead to 1.5% turnover savings from preventing fraud operations.

 Embeddings analysis & finetuning of BERT models in language modelling and its application within sentiment analysis task. Step by step finetuning for the text data was done: - sequence classification model was studied on classification task embeddings with usage of classes construction that was received from mlm & sentence prediction task;- mlm transformer embeddings were used in classification transformers.  NER (conditional random field & classical models with features generation and augmentation). Exploratory analysis & primary text processing: top words identification & TSNE visualisation; text statistics; synonyms, associations & doesn’t match words on the basis of word2vec, FastText & CNN classification.

Key Results & Achievements: Workout of customer profile & decision criteria in Moscow Region townhouse purchase process on the basis of CRM system analysis (Feature importance of LR, DT, RFC models). Applied experience of convertation of business task to algorithms one and ML models practical implementation & support. Explanation of ML algorithms to people without math & machine learning knowledge.Business effect: Deeper client need understanding helped in product advertising approach correction & 2% sales increase Underlining of key product features from client perspective helped in more effective company budget distribution and cost decrease by 3%. More targeted advertising based on client need identification lead to increase of newcomers by 2%.

Key Results & Achievements:  Development of analytical methodology to ensure automatic & mass court cases resolution led to reduction of investors direct requests to court system in 2 times and developer monitoring procedure length from 10 years to 15 months in group of 4 people. Main Functions: • Investors status analysis process (20 ths people, 6 ths agreements). Preparation of mass investors inclusion.

Key Results & Achievements:  Strong knowledge of finance algorithms (fin. controls, management reporting) and experience of their implementation in business supported local management team in proactive company KPIs management and led to increase in forecast accuracy by 20% & team performance bonus increase by 15%. Detailed cash flow planning, credit limits installment, working out effective credit control procedures and negotiations with leading intergrators led to cash flow growth by 80%. Finance topics explanation to non-finance team & convertation of business-tasks to finance structures helped to create management reporting packs. Main Functions:• F&A management, business reviews, KPIs improvement, IFRS statements, financial control, management reporting, budgeting & forecasting, IT system improvement, processes organization, people management.

Key Results & Achievements:  Middle management reporting packs work out & financial modelling (products, new plant) on the basis of large data volumes analysis led to increase in forecast reliability by 30% & better decision-making. Explanation of finance models to non-finance team led to higher finance culture & business adaptability in crisis time and helped in 5% sales and 3% cost economy increase through more effective budgeting process.  Strong knowledge of finance algorithms and experience of their implementation in business were applied to strategic projects enforcement: 7 material risks were identified and mitigated after 40 SOX controls implementation; leading finance part of BP re-engineering help to correct 30 duplicated processes & reduce processes inefficiency. Main Functions:• F&A management, business reviews, finance control, budgeting & forecasting, BP re-engineering, projects management, IT system post-implementation, processes automation. Financial models & algorithms workout & practical business application. Convertation of business tasks to finance ones. Data analysis.