1. Developed an in-house data quality framework utilizing Pydeequ, SQL, and Pyspark in Databricks.2. Implemented comprehensive data quality checks at various stages of the ETL process, aligning them with business requirements and understanding 3. Collaborated with downstream stakeholders to identify recurring issues in data products, leveraging their insights to define effective data quality checks4. Integrated data quality checks as dedicated tasks within the existing Airflow DAGs of the ETL pipeline, enhancing the overall data quality assurance process5. Analyzed the impact of duplicate values on key performance indicators (KPIs) and implemented solutions to mitigate their effects.6. Conducted an analysis of merchant-level information across multiple datasets, enhancing the depth and accuracy of our data products 7. Utilized Google Analytics session data and leveraged Google BigQuery to assess the reliability of merchant information, ensuring high data quality and integrity8. Designed and implemented the ETL batch processing of terabytes of weekly data to improve the accuracy of the complex KPIs - Traffic & Demand weighted availability rate and Adjusted rate of sales.9. Identified and resolved existing bugs in the data product, resulting in improved KPIs by enhancing the ETL using Airflow and Databricks10. Migrated GA pipeline from Oracle to cloud-based storage platform Amazon S3 using Airflow, Pyspark, and Databricks

Quantitative and qualitative evaluation of machine learning and deep learning classification models in NLP with a focus on the automobile domain - Interpretability, Robustness, and Fiarness of BERT LM

1. Literature survey on the parameters used for the probability of default of a company2. Trained the neural network MLP model using Tensorflow framework and Keras Tuner for hyperparameter tuning and model explainability using SHAP and LIME3. Information retrieval of financial data about the company from the recent news

1. Literature survey on the parameters used for corporate credit risk rating2. Feature Engineering - extracting financial KPIs, cleaning, and organizing the data3. Visualization of different KPIs of companies using python - seaborn, matplotlib, and plotly4. Experimenting with different machine learning models (Random Forest, XGBoost, CatBoost and Neural Networks) with hyperparameter tuning to best classify the corporate credit risk rating5. Built an Excel front-end model to take financial parameters as input and output the rating predicted by model

1. Experience in the support and enhancements of the four applications2. Built ETL pipeline using Informatica power center to extract the data from multiple sources such as flat files, XML files, databases like Netezza and oracle 10g to staging area and finally to dataMart3. Created mappings using different transformations, Informatica source analyzer, and lookups4. Resolving issues in real-time scenarios5. Experience in writing complex SQL queries for analysis and extraction6. Performed impact analysis of the changes done to the existing mappings7. Worked on automation of the application to avoid manual errors and intervention