Meet Vasani Email & Phone Number

Cleveland, Oh, Us

Jersey City, New Jersey, United States

● Automated repetitive data extraction tasks from Snowflake database using SSIS, and streamlined data collection process● Applied data transformation, feature selection, regularization and bias-variance trade-off using Python libraries such as Pandas, NumPy, and scikit-learn reducing model complexity by 5%● Performed hyperparameter tuning, integrating machine learning models, automating failure detection process, energy consumption trends, improved resource allocation and reducing error by 10%● Engineered and optimized scalable data pipelines using dbt and AWS for data transformation, git for version control, SSAS with predictive analysis, advanced SQL and OLAP cubes for multidimensional analysis, reducing processing time by 8%● Integrated Power Automate with Power BI to establish real-time data updates in dashboards, enhancing KPI tracking accuracy and facilitating decision-making for stakeholders

India

● Conducted advanced data preprocessing techniques, including extraction from Postgres, data cleaning, feature engineering, and outlier detection, ensuring high-quality datasets for model training, improving overall data reliability by 12%● Demonstrated robust data practices including A/B testing, funnel analysis, customer churn analysis, user behavior tracking, optimizing strategies through feature selection, yielding 15% improved ad spend efficiency● Leveraged advanced SQL CTEs, window functions, and subqueries, alongside Tableau to create dashboards, and producing ad-hoc reports using SSRS, supporting merchandising strategies and boosting 10% overall campaign performance● Spearheaded cross team collaboration using Python, conducting tasks like imputation, normalization, and scaling, resulting in 20% faster data analysis and model development● Methodically segregated and compared baseline, ensemble and deep learning models via various metrics (i.e., RMSE value, MSLE value and R2 score), achieving a substantial reduction in error rates and delivering more accurate predictions

India

● Innovated advanced Gujarati Question Answering system utilizing machine learning technique for data collection, data – preprocessing, and establishment of relation between user for precise answer extraction from provided content● Optimized question processing by 5% accuracy via pre-processing, IndoWordNet API tokenization, bag of words models, count vectorizers and NLTK for accurate answer retrieval, catering to diverse question structures● Performed EDA using Python on public sentiments and 11 emotions related to COVID 19 using 5000 tweets from Twitter● Excelled in Sentiment Analysis using SVM with TF and TF-IDF achieving high accuracy, precision, recall, and F1-Score, also considering Emotion Detection where SVM outperformed in TF-IDF, while MNB, XGBoost, and CNN surpassed in TF

Master Of Science - Ms, Data Analytics Engineering

Courses: Fundamentals of Data Analytics Data Management for Analytics Data Mining in Engineering Computation and Visualization.

Bachelor Of Technology - Btech, Computer Engineering

Co-authored 3 International research papers in the field of Data Science/ Analytics, (Springer)