Spondon Basu

About Spondon Basu

Seasoned IT professional with 11+ years of experience, including over 5 years in Data…

Experience

1. Specialist

   LTIMindtree

   Sep 2021 — Present · Kolkata, IN

   PROJECTS UNDERTAKEN: • Crack Detection in Concrete Pavement Using Images (Convolutional Neural Networks) • Developed a Convolutional Neural Network (CNN) model using TensorFlow to detect cracks in concrete images. • Achieved an accuracy of 92.3% in identifying cracked and non-cracked concrete surfaces. • Revenue Forecasting with Time Series Analysis (ARIMA Model) • Analyzed sales data using ARIMA models to forecast revenue. • Decomposed the time series to identify seasonality patterns and used Grid Search and Forward Walk validation to optimize the model. TECHNOLOGIES: Python, Scikit-learn, NumPy, Pandas, Feature Engineering, Machine Learning, Predictive Modeling, Time Series Analysis, TensorFlow, Computer Vision, Image Processing, Deep Learning, Convolutional Neural Networks (CNN) RESPONSIBILITIES: • Data Preprocessing and Feature Engineering: Preprocessed and feature-engineered large datasets using Python, NumPy, Pandas, and scikit-learn. • Model Development and Deployment: Designed, developed, and deployed machine learning models, including a CNN using TensorFlow and an ARIMA model for revenue forecasting. • Performance Evaluation and Optimization: Evaluated model performance using metrics such as accuracy and mean absolute error. Employed techniques like Grid Search and Forward Walk validation to optimize models. • Collaboration: Collaborated with teams to integrate models into existing workflows.