I'm a data scientist who is passionate about extracting insights from data, building models, and developing data-driven solutions. My work spans across machine learning, deep learning, natural language processing and data visualization. I enjoy contributing to open-source projects and collaborating with the data science community.
| 📝 Project Name | 🔍 Description | 🛠️ Tech Stack | 📎 Links |
|---|---|---|---|
 Credit Card Fraud Detection |
A RandomForestClassifier model to predict credit card fraud using credit card data from Kaggle. Achieved 99% accuracy. | Python, scikit-learn, Matplotlib, seaborn | Repo |
 Fashion Data Classification |
Developed a Convolutional Neural Network (CNN) using Conv2D layers to classify images from the Fashion MNIST dataset. The model accurately distinguishes between 10 categories of clothing and accessories by extracting spatial features from images. Optimized for performance, the network achieves high accuracy in identifying various fashion items. | Python, Tensorflow, Keras, scikit-learn, Matplotlib, seaborn | Repo |
 Chest X-ray Analysis |
Performed analysis on chest X-ray data to detect abnormalities, leveraging advanced machine learning techniques. Visualized results using heatmaps and feature maps to highlight key areas of interest in the X-ray images. | Python, Pandas, NumPy, Tensorflow, scikit-learn, Matplotlib, seaborn | Repo |
 Text to Speech |
Leveraged the pretrained Bark model to generate high-quality and natural-sounding speech from text inputs. The model was fine-tuned to handle various accents, tones, and languages, enabling robust text-to-speech synthesis across different domains. The approach includes optimizing the model for fast inference, ensuring high performance for real-time speech generation applications. | Python, TensorFlow | Repo |
 Hungarian Chickenpox Prediction |
Developed a predictive model utilizing Bidirectional LSTM (Long Short-Term Memory) to forecast chickenpox cases in Hungary. The model was trained on a publicly available dataset sourced from the UCI Machine Learning Repository, Hungarian Chickenpox DatasetDataset reference: Hungarian Chickenpox Cases | Python | Repo |
 Cover letter generator |
Cover letter generator is used to generator a standard cover letter. | Python | Repo |
- Theses 1: Text Professionalisation using BART Model.
- Blog 1: Harnessing Machine Learning for Heart Attack Awareness: A Deep Dive with Visual Insights
- Blog 2: Unleashing the Power of Voice: Exploring BARK’s Free Text to Speech & Voice Cloning Capabilities
- Blog 3: Harnessing Machine Learning to Filter Emails: A Naive Bayes Approach
- Blog 4: Harnessing the Power of AI: Revolutionising Cybersecurity in the Digital Age
