An innovative approach of integrating both text embeddings and noise vectors to generate realistic images based on textual descriptions using Deep Convolutional Generative Adversarial Networks.

A research project in optimizing the IREE ML compiler. Achieved an up to 70% performance uplift for the VMVX runtime target on the Intel Golden Cove µ-Architecture.

A heuristic approach of finding pick-up locations that optimizes the transportation plan for the Party Together problem, which is an NP-Hard problem, using a genetic algorithm.

Enhancing machine learning models with natural parallelism using OpenACC. Accelerates softmax regression, neural networks, and CNNs by optimizing for GPU parallel computing.

Adapting sorting algorithms for parallel execution using MPI. Focuses on task division, inter-process communication, and synchronization for improved performance.

An AI assistant project that leverages the analytical power of GPT models within the realm of Natural Language Processing. Aimed at creating seamless human-computer interactions, this platform showcases the integration of conversational AI with NLP for enhanced user experiences.

Optimizing matrix multiplication efficiency through natural parallelism by leveraging advanced memory management, data-level parallelism with SIMD, and multi-threading with OpenMP and MPI to enhance computational speed and scalability.

An advanced exploration of image processing utilizing the power of natural parallelism. Designed to enhance computational efficiency, showcasing the potential of parallel techniques in real-world applications.

My website featuring insightful blogs and a portfolio of my projects.

A design and implementation of comprehensive banking database. Built using SQL, Python, HTML, CSS, and JavaScript.

An implementation of unsupervised learning algorithm: K-means Clustering. Built using HTML, CSS, and JavaScript.

An application of Multilayer Perceptron on recognizing ten thousands of handwritten digits. Built using Python.

A simple application of linear regression on predicting next-day temperature. Built using Python.

A machine learning model designed to classify emails as either spam or legitimate. Achieved an accuracy of 85.7%. Built using Python.

A Snake game with an AI Monster feature. Built using Python.