Alexandra Joukova

I hold graduate degrees in both data science and economics, bolstered by over seven years of experience in designing, implementing, and evaluating predictive models. This journey encompasses the entire quantitative model development cycle, from data exploration and cleaning, feature selection, to methodology development, implementation, documentation, defending, and execution. Specifically, I have created regression models (linear, logistic, beta) to forecast the loss given default and the probability of default, followed by an analysis of coefficients and residuals and model evaluation. The portfolios I've handled include both wholesale and retail data, involving revolving credit lines and auto loans. This demonstrates my capacity to manage and analyze a diverse range of data types and navigate the entire model development cycle.

My expertise extends to over four years in Natural Language Processing (NLP), where I employed semantic similarity and topic modeling for unsupervised categorization of free text, focusing on ngrams, Named Entity Recognition (NER), fuzzy matching, text summarization, and topic modeling. For instance, I used NER for the extraction of human names from free text and built a neural network to detect whether strings are human names. This specialization is further enriched by over two years of experience in evaluating marketing campaign strategies using statistical methods, and a solid background in machine learning, encompassing decision trees, neural networks, and clustering.

I have utilized neural networks, such as random forest, for predicting depression diagnoses and patron casino visits. Besides, I am proficient in SQL, Python, and R, and have garnered extensive experience in visualization using PowerBI, creating metric dashboards, patron statistics reports, and patron networks (graphs). This proficiency extends to converting large, complex data sets into actionable insights, refined by a deep understanding of probability and statistics, experimental design, machine learning, data mining, and predictive modeling. This comprehensive skill set ensures a robust analytical approach and the ability to translate these analyses into impactful business strategies. Additionally, my thorough understanding of statistical analysis and experimentation design, combined with proficiency in common analytical techniques like regression and decision trees, equips me to approach problems from a well-rounded, thoughtful perspective.

I also have experience managing data pipelines in Azure Synapse and adeptly utilizing PySpark notebooks. In Azure ML, I have managed data pipelines, executed script updates, and adjusted pipelines. Moreover, I am proficient in version control and managing scripts using Git. This array of skills ensures robust data management and analysis, as well as effective version control and pipeline management.

7+ years of experience in design, implementation and evaluation of predictive models
4+ years of experience in NLP - ngrams, Named Entity Recognition (NER), fuzzy matching, text summarization, topic modeling
2+ years of experience in marketing campaign strategy evaluation with statistical methods
Strong background in machine learning, including decision trees, neural networks, and clustering
Developed business insights through client segment analysis and consumer behaviour research
Improved sales and budget forecasting accuracy through regression and time series analysis
Designed reports and performed data visualization using Power BI and graphing capabilities of SAS, R, and Python


Python: experienced in data science stack, including spaCy for named entity recognition, pandas, numpy, and scikit-learn for machine learning and data manipulation, and keras for building and training deep learning models
SAS and R: performed regression analysis, predictive model development, decision tree analysis
Power BI: created and customized dashboards, with proficiency in using DAX functions and writing queries
MS SQL Server, MYSQL workbench: hands-on experience
Analyzed datasets using Pig, Hive, knowledge of PySpark and MapReduce implementation
Azure:
Azure Synapse: Managed data pipelines and adeptly utilized PySpark notebooks.
Azure ML: Ran data pipelines, executed script updates, and adjusted pipelines.
Version Control: Proficient in controlling and managing scripts using Git.

I hold a Master of Science in Data Science and Analytics from Toronto Metropolitan University, during which I published a paper titled "Social influence-based contrast language analysis framework for clinical decision support systems". Additionally, I completed PhD-level economics courses at McMaster University and published another paper, "Visualizing data: Trends in smoking tobacco prices and taxes in India". I also possess a Master of Arts in Economics from Waterloo University. This educational background equips me with a deep and comprehensive understanding of both the data science and economics fields, as evidenced by my published works and advanced studies.