- 🎯 Currently pursuing: A Master of Science (MSc) in Data Science from the University of Pittsburgh, and harnessing the power of Computer Vision to develop innovative solutions that address complex, real-world problems. Leveraging cutting-edge technologies to drive impactful results and foster innovation in the field of AI.
- 📚 Continuing Learning: Immersing myself in advanced deep learning algorithms and their applications across various domains. Always striving to stay at the forefront of technological advancements and best practices in the industry.
- 🚀 Passions: Ardently dedicated to Data Science, Machine Learning, Deep Learning, and Computer Vision. Eager to collaborate, share insights and discuss groundbreaking ideas that push the boundaries of what’s possible.
- ⚡ Fun fact: In the realm of data and algorithms, "Impossible" is simply "I'm-Possible", yet to be decoded.
| Title | Description |
|---|---|
| DishVision: Multi Class Image Classifier - Computer Vision | End-to-end project utilizing PyTorch for deep learning and computer vision, classifying images of three different types of food with 96+% accuracy on the test set. Includes a detailed replication of a Vision Transformer machine learning research paper. Deployed on Hugging Face as an interactive app and can be accessed here. |
| Food Image Deca (10 Classes) Classifier - Computer Vision | Classifying images of 10 different food classes using deep learning and computer vision techniques with TensorFlow. Achieved 90% accuracy on the full test set. |
| Food Image 101 Classes Classifier - Computer Vision | Leveraging TensorFlow for transfer learning and fine-tuning, this project classifies 101 different types of food with 80% accuracy on the test set. This performance outperforms the 50.76% baseline accuracy reported in the original Food101 paper using 10% of the training data and surpasses the 77.4% accuracy benchmark set by the DeepFood paper using 100% of the training data. |
| BTC Price Predictor - Time Series | Time series forecasting project aimed at predicting the price of Bitcoin using machine learning and deep learning techniques with TensorFlow. |
| Smart Abstract Classifier - NLP | Developing a Natural Language Processing model to classify sentences in abstracts, articles, and other texts into their respective roles, using TensorFlow. |
| HireMind - NLP | A private, in-progress job management web-app merging advanced ML and DL models, NLP techniques, and robust management. |
| Webcam Guardian Alert - Computer Vision (OpenCV) | A monitoring tool that uses your webcam to detect movement and send email alerts. |
| PropPredict | Developing a Machine Learning Real Estate Price Prediction Engine. |
| Title | Description |
|---|---|
| MeteoMate 🌦️ | A Flask-powered weather web-app built with Python, Bash & JavaScript, serving real-time weather data for select cities. Features responsive UI, Docker containers (for the app and its unit tests), and CI/CD pipelines via GitHub Actions. Deployed on Heroku and AWS EKS using Terraform, Kubernetes, and Helm, showcasing continuous evolution. • View live on Heroku: MeteoMate on Heroku • View live on AWS: MeteoMate on AWS |
| Title | Description |
|---|---|
| GlobalForecast5 | Web application providing weather forecasting for any place globally for up to 5 days, visualized using line graphs or relevant images. |
| WeatherRESTful API | Flask-based web application providing access to temperature data from various weather stations around the world. |
| CSVPDFusion | Python-based solution for generating PDF files for note-taking on specific programming topics, sourced from a CSV file. |
| Sheet2Receipt | Python application that converts invoice data from Excel files into structured PDF documents. |
| EventEagleEye | Python-based application designed to periodically scan a specific website for new tour event announcements and send email notifications. |
I regularly post detailed articles on LinkedIn, discussing various topics in data science, machine learning & deep learning. You can explore my posts and connect with me on LinkedIn:
Feel free to explore my work and connect with me through any of the platforms listed above. Let's exchange knowledge and build something amazing together 🚀