• Developed and implemented a forecasting model to predict the bank's liquidity, enabling the finance department to generate accurate future liquidity reports and prevent a 5% penalty due to inaccurate liquidity projections. • Conducted in-depth analysis of deposit flows to track customer funds entering and exiting the bank, creating a comprehensive dashboard that provided the bank with a holistic view of customer behaviors and trends. • Utilized clustering techniques to analyze and rank over 300 bank branches based on performance metrics, supporting the planning department in setting precise, data-driven goals for branch-level improvements. • Facilitated over 30 meetings with stakeholders, presenting project updates and adapting project scope based on feedback to align deliverables with the bank's strategic objectives.

RCDAT is a research center that does research-industrial projects for various companies and enterprises. Iwas in an Image Processing team. • Implemented a Faster RCNN-based model to detect Integrated Circuits (ICs) on electrical boards, achieving a 93% accuracy on the dataset. • Implemented a YOLOv5-based model for detecting backpack, achieving a 88% accuracy on the dataset. • Utilizing image processing techniques for required dataset preprocessing and post-processing.

NABZ is a leading provider of smart health gadgets in Iran. • Implemented an algorithm to extract the respiration rate from ECG, achieving a precision of ±1.0 respiration per minute. • Implemented an algorithm to detect the onset of the T wave in ECG with a precision of ±1 millisecond. • Implement a neural network model, incorporating CNN, LSTM, and fully connected layers, to estimate blood pressure from ECG and PPG data.