Machine Learning and Deep Learning enthusiast with a solid CS background and hands-on experience in building end-to-end intelligent solutions using Python, TensorFlow, PyTorch, and Scikit-learn. Skilled in computer vision, NLP, and predictive analytics, with expertise in model tuning, evaluation, and deployment using Streamlit, Flask, and Docker. Passionate about AI research, model development, and impactful innovation.
0 Projects completed
My academic journey in Computer Science and hands-on experience in Machine Learning and AI-related initiatives.
MAKAUT, West Bengal | CGPA: 8.4
Relevant coursework: Data Structures, Algorithms, AI, Machine
Learning, DBMS
WBCHSE Board |Percentage: 89% | PCMB
WBCHSE Board | Percentage: 89.71%
Developed a sentiment classifier using TF-IDF and Random Forest, deployed via Streamlit for restaurant review system.
Built a Streamlit app for image classification using a custom CNN and MobileNetV2 on CIFAR-10, enabling model selection and live predictions
Technical skills and tools I've mastered in Machine Learning and Software Development
C
C++
Python
Java
MySQL
PyTorch
Scikit-learn
NumPy
Pandas
Matplotlib
Seaborn
OpenCV
Azure
Docker
VS Code
Jupyter
Colab
Power BI
Selected machine learning projects demonstrating my technical capabilities
Attention-Enhanced GCN with SPP for UAV-Captured Plant Imagery
Engineered an attention-augmented Graph Convolutional
Network (GCN) integrating EfficientNet-B0, ShuffleNetV2, and
Spatial Pyramid Pooling (SPP) for real-time UAV-based plant
disease identification..
Designed a hybrid architecture featuring multi-branch
spatial feature fusion and graph-based relational reasoning
for robust aerial image understanding.
Sentiment Analysis of Restaurant Reviews
Developed a sentiment classifier using TF-IDF and Random
Forest, deployed via Streamlit for restaurant review
analysis with 85.5% accuracy.
Streamlined customer feedback workflows by automating
sentiment insights, reducing the need for manual processing.
Image Classification Web Application
Built a Streamlit app for image classification using a
custom CNN and MobileNetV2 on CIFAR-10, enabling model
selection and live predictions, achieving 84.77%
accuracy.
Developed an interactive interface supporting instantaneous
inference, enhancing accessibility for non-technical users
in practical scenarios.
Feel free to reach out for collaborations or opportunities
Durgapur, West Bengal, India