End-to-end machine learning project applied to real-world business problem. This project is trying to depict rental situation in Toronto and find out what affects the current prices. Focus is on exploratory data analysis, feature engineering and hyper parameter tuning.
Full-Stack Geospatial Web Application integrating Django backend with RESTful APIs and Three.js frontend. It enables dynamic geospatial data visualization on a globe, with a focus on server-side data management and responsive front-end interaction, utilizing Geocoder API and RESTful services for efficient data handling.
Leveraging the strengths of BERT and Pegasus models, I developed an advanced text summarization system that enhances NLP performance. I streamlined deployment with CI/CD pipelines on AWS, encapsulating the model within Docker containers for consistent operation across various environments. Throughout the project, I adhered to agile methodologies to ensure scalability and incorporated the latest advancements in AI to meet evolving needs.
I engineered a sophisticated ETL pipeline tailored for the Tokyo Olympic data, harnessing Azure Data Factory for efficient data extraction and loading. By utilizing Spark on Azure Databricks, I optimized the data transformation process and enhanced storage in Azure Data Lake. Additionally, I created dynamic, interactive dashboards in Power BI, driven by Azure Synapse Analytics, to provide comprehensive insights for data-driven decision-making.
Applied cutting-edge tools like GCP Storage, Mage Data Pipeline Tool, BigQuery, and Power BI to analyze Uber data, showcasing expertise in handling large-scale datasets and driving data-driven decision-making.
