Greetings! I'm Sai Shirini, a highly analytical individual and a Master's student pursuing a degree in Data Science at the University at Buffalo, The State University of New York. With a passion for continuous learning and collaborative problem-solving, I am excited to embark on a challenging career in the field of data science. This portfolio showcases my work in Data Science including projects I've completed for academic and self-learning purposes.
- Email: [email protected]
- LinkedIn: linkedin.com/sai-shirini
- M.S. in Data Science (Expected Dec 2024) - University at Buffalo, SUNY
- B.Tech in Electrical and Electronics Engineering - Mahindra Ecole Centrale, Telangana
- Research Assistant - University at Buffalo : Actively engaged in continuous research and development, I am currently working on PyTorch-based neural network models for precise video event detection. Utilizing advanced architectures, temporal modeling techniques, and continual learning, my goal is to enhance model accuracy.
- Project Researcher - IIT Palakkad : I actively contributed to the advancement of computer vision through in-depth research on Justifiable Convolutional Neural Networks (J-CNN) and Grad-CAM. My primary focus was on implementing the J-CNN model to enhance the interpretability of Convolutional Neural Networks, achieving a notable precision of 0.84 in a binary classification task.
- Grader - IIT Palakkad : I served as a Grader at IIT Palakkad, collaborating closely with professors to oversee the compilation and assessment of class assignments for a student cohort of 150. This role demanded a high level of organization, attention to detail, and adherence to academic standards.
This project demonstrates my application of deep learning techniques for predicting action labels in videos. Employing various key-frame extraction methods, I focused on extracting pivotal frames to enhance the accuracy of the classification model.
This project showcases my adept application of deep learning methodologies, emphasizing the use of advanced techniques to enhance sentiment classification performance. Sentiment analysis was conducted on the Sentiment140 dataset employing neural networks, with a notable achievement of 0.769 accuracy utilizing Bidirectional LSTM.
I employed advanced machine learning algorithms to develop a predictive model for anticipating the rise in COVID-19 cases, achieving an impressive accuracy rate of 88%.
I implemented an LSTM neural network for stock price prediction, utilizing deep learning to forecast prices based on historical data.
In this project I Analyzing different types of incidents in Buffalo is crucial for understanding the city's safety landscape. Sorting and organizing incidents into groups can provide valuable insights for law enforcement and other safety-related organizations.
In this project, I delved into the realm of image processing, exploring optimized partial differential equation (PDE) models for image filtering using Matlab.
- Languages: Python, R, SQL, Scala(Basics)
- Machine Learning: Supervised/Unsupervised learning, Feature selection/extraction, Ensemble methods, Model evaluation/selection
- Deep Learning: Feed Forward Neural Networks, CNN, RNN, LSTM, GAN, Transformers
- Frameworks: Pandas, Matplotlib, Scikit-learn, TensorFlow, PyTorch, OpenCV, NLTK
- Data Science: Data preprocessing, Data mining, Exploratory data analysis, Natural Language Processing, Computer Vision
- Others: PowerBI, Matlab, VSCode, RStudio, Colab, Jupyter, MySQL, PostgreSQL, Hadoop, Spark
- Secured All India Rank 77 in GATE Mathematics'22
- Secured All India Rank 554 in JAM Mathematics'22
- Received Rs.1 lakh as merit scholar ship from Mahindra Ecole Centrale in 2018
This portfolio undergoes regular updates to showcase new projects, achievements, and certificates, reflecting my ongoing professional growth.
For inquiries or professional connections, you are welcome to reach out via email or connect with me on LinkedIn!