• Developed an AI Platform application using FastAPI to host ML models in production. The application hosted model structures with low compute resources, achieved through threading and parallel computing resulting in accuracy of 99% in classifying parcel warehouses based on address entered by the user.• Implemented various OCR strategies to extract data from bank transaction slips using PyTesseract and EasyOCR which resulted in saving user’s time by 40%.• Designed and implemented various features in a Financial Application for Expense Management, utilized company-wide. Suppliers, vendors, and employees use this tool to track company expenses at a warehouse level, resulting in improvement in expense tracking efficiency.

• Designed and Developed Attendance Management System in Golang along with AWS Rekognition which resulted in managing employee attendance efficiently and saving employees time for marking attendance by 75%.• Designed and developed a Python-based Risk Management application using GRPC and Markov Models. Implemented to identify and validate addresses in Bangladesh, automating data validation efforts with a 75% accuracy improvement.• Designed and implemented various features in a Financial Application.

• Developed various object detection models achieving an total loss of 0.25 in identifying ripe sweet oranges from trees.• Labeled collection of Sweet Orange image data (250 images) manually gathered from a farm using Labelling Tools.• Implemented machine learning models using Python and TensorFlow, boosting accuracy of object detection by achieving low classification score of 0.08, demonstrating hands-on expertise in algorithm development and data preprocessing.