During my role as a Research Assistant at Western University, I leveraged my skills in backend and frontend development using Python and Java to create software solutions for research in robotic automation and intelligent transportation systems. This effort led to a 30% improvement in system efficiency. I conducted comprehensive literature reviews on autonomous mobile robotics and transportation, ensuring our projects were grounded in the latest advancements. A significant highlight was developing a customized ChatGPT-based indoor navigation system, which reduced user guidance errors by 25% and provided real-time assistance for indoor environments in SMART CITY initiatives. Additionally, I collaborated with a multidisciplinary team to design and prototype solutions that seamlessly integrated robotics, AI, and IoT technologies, resulting in a 15% improvement in project turnaround time. This experience not only enhanced my technical and analytical abilities but also honed my skills in cross-functional teamwork and innovative problem-solving.

I developed a comprehensive Financial Market Analysis and Sentiment Suite, integrating advanced data extraction techniques to enhance market response analysis accuracy by 20%. By implementing sentiment analysis tools using Python and PyTorch, I improved sentiment accuracy by 30%, enabling more precise market predictions. Additionally, I incorporated Google Trends data to strengthen the correlation between market trends and sentiment analysis, achieving a 20% improvement in predictive capabilities. Leveraging multiple APIs, I designed a robust framework that streamlined data processing and analysis, contributing to more informed financial decision-making. This project not only deepened my expertise in financial data analytics but also reinforced my proficiency in machine learning and real-time data integration.

I designed and implemented a speech recognition and NLP system using PyTorch, enabling accurate audio-to-text conversion, feature extraction, and intelligent response generation. By integrating deep learning models such as CNNs, RNNs, and Transformers trained on the LibriSpeech dataset, the system achieved over 90% accuracy in speech recognition and 80%+ accuracy in NLP tasks. Performance optimization was a key focus, where I applied hyperparameter tuning, regularization, and transfer learning with pre-trained models to enhance efficiency and robustness. Utilizing Python and multiple APIs, I developed a scalable solution that demonstrated strong real-world applicability in voice-based AI interactions. This experience strengthened my expertise in deep learning, natural language processing, and speech recognition technologies.