Developing and implementing data quality assurance frameworks to ensure accuracy, completeness, and consistency of large-scale interior sensing camera datasets for AI/ML applications.

Designing robust validation processes to detect and correct data anomalies, inconsistencies, and errors in image and sensor-based data.

Building tools and metrics to assess image quality, accounting for variables such as lighting conditions, object diversity, and sensor calibration to meet AI/ML model training standards.

Validating ground truth annotations and sensor-derived data to ensure reliability for supervised model training and evaluation.

Built a prescreening system to identify genuine individuals by assigning probability scores across domains such as money laundering and politically exposed persons using a large-scale domain-specific dataset.

Applied NLP techniques for semantic string matching, involving data cleaning, preprocessing, model building, and evaluation.

Implemented NLP models using libraries like spaCy, NLTK, and scikit-learn, Word2Vec, Weighted Average Word2Vec, and TF-IDF to optimize prediction accuracy.

Conducted statistical analysis to evaluate model performance using accuracy, F1-score, recall, and confusion matrix.

Deployed the service through a CI/CD pipeline using AWS, Docker, and Kubernetes for scalable and reliable production delivery.

Built a financial compliance tool using Django Admin to automate daily ingestion of Internal Quality Control (IQC) and duplicate data from Excel/CSV files into SQL Server, eliminating manual data updates.

Implemented multi-select filters, advanced search, and user authentication features (signup/login/logout) to enhance usability and access control.

Designed and integrated custom PDF report generation using ReportLab, allowing dynamic content creation based on user input.

Enabled a data import/export functionality and deployed the application using Nginx with MongoDB for flexible data storage and querying.

Developed a tool to automatically parse unread job alert emails, extract relevant HTML content, and filter out unnecessary information based on configurable rules.

Structured and exported parsed data to various formats including PDF, DOCX, and CSV for easier consumption and analysis.

Designed a back-end service to process the extracted email data and generate structured outputs based on user-defined criteria.