Conducted research to develop best practices for the operational aspects of Large Language Models, focusing on cost-efficiency, scalability, and performance optimization within LLMOps frameworks.

This project has been focused on exploring the Retrieval-Augmented Generation (RAG) in large language models.
The mechanism, advantages, applications, and challenges of RAG, along with its potential future developments and impact on AI and machine learning, have been examined. This exploration is essential for understanding the evolving role of AI in various sectors.

Utilized OpenAI's Large Language Model (LLM), employing prompt engineering with no manual coding to create my portfolio.
Successfully built an interactive three-page portfolio website, showcasing the transformative potential of LLMs in practical applications.

Demonstrated how GAN’s architecture can be used in style transfer. Showcased the generalizability and adaptability of CycleGAN.
Implemented image-to-image translation GAN using TensorFlow packages to generate artistic paintings from real-world photos.

Devised ensemble methods (ADA-Boost, Gradient Boost, and Extreme GB) with 98% accuracy.
Gained in-depth insights into various boosting and bagging algorithms.

Determined traffic volume of Minnesota through deep learning algorithms, including CNN, RNN, and LSTM.
Evaluated various deep neural network algorithms using Keras packages to provide accurate time series prediction.

Collaborated with a team to conduct data analysis on the HR dataset and create insightful visualizations.
Utilized Tableau and R to explore relations between employees' age, gender, department, and tenure in the workplace.

Executed different ML algorithms for classifying Intel Image dataset, attaining high accuracy rates.
Demonstrated expertise in optimizing CNN models which improved accuracy by 12%.

Analyzed the cost of homes in Germany through principal component analysis (PCA), regression, and clustering methods.
Conceptualized and designed a highly effective model with a random forest learning method to build a recommendation system.