Projects

Customer Transactions & Loan Data Pipeline Using Azure Databricks

Built an end-to-end data pipeline to process transactions and loan data across Bronze, Silver, and Gold layers in ADLS Gen2.
Used Databricks Notebooks for cleaning, joining, and transforming five key datasets in the Silver layer.
Applied deduplication and implemented SCD Type 1 logic for structured, analytics-ready output.
Stored final data in Delta format and used Azure Key Vault for secure configuration management.
Visualized insights through Power BI dashboards and published reports to Fabric Workspace.
Technology Used: Azure Databricks, ADLS Gen2, Delta Lake, Parquet, Azure Key Vault, Power BI, Fabric Workspace.

Designed a complete data pipeline to process customer account data using Azure Data Factory (ADF) and ADLS Gen2.
Ingested and organized raw data from backend storage into the Bronze layer for structured processing.
Implemented data cleaning and transformation using ADF Dataflows, ensuring high data quality.
Applied Slowly Changing Dimension (SCD) Type 1 and Type 2 techniques for accurate historical data management.
Stored processed data into Azure SQL Database and visualized insights through Power BI dashboards.
Automated the entire pipeline with scheduled triggers to support continuous, scalable analytics.
Technology Used: Azure Data Factory, Azure Data Lake Storage Gen2, Azure SQL Database, Power BI, ADF Dataflows (SCD Type 1 & 2), and Delta/Parquet formats.

Developed an AI tool that predicts the next relevant word in chat applications.
Utilized natural language processing, language modelling, and deep learning techniques.
Implemented features like text tokenization, word frequency analysis, and a confidence score system.
Designed a user interface for seamless interaction and integrated API for broader application use.
Aimed to enhance communication by reducing language barriers, saving time, and preventing spelling mistakes.
Technology Used: Natural Language Processing, Deep Learning, Python

Developed an automated rule-based system to streamline Canadian tax filing.
Simplified complex tax computations using CRA guidelines for accuracy.
Designed an intuitive user interface for easy data entry and navigation.
Integrated a knowledge base of tax regulations ensuring up-to-date computations.
Validated system outputs against expert-prepared returns for reliability.
Technology Used: HTML, Flask, Python, CRA Tax Regulations

Developed a chatbot for real-time 24/7 appointment scheduling in healthcare and professional services.
Utilized natural language processing to understand user requests and machine learning to optimize appointments.
Designed to enhance user satisfaction, reduce administrative workload, and improve operational efficiency.
Integrated advanced features for seamless user interaction and instant booking confirmation.
Aimed at creating a more organized and customer-friendly reception system.
Technology Used: Natural Language Processing, Machine Learning

Conducted exploratory data analysis using Python libraries such as Pandas, Numpy, and Matplotlib.
Implemented an LSTM-based model to predict Bitcoin price movements.
Developed a trading strategy to optimize buy/sell decisions for profitability.
Focused on short-term price predictions to inform trading actions.
Aimed to capitalize on cryptocurrency market trends for financial gain.
Technology Used: Python, LSTM, Pandas, Numpy, Matplotlib