• Conducted research on deep learning and transfer learning, designed, and implemented models integrating transfer learning methodologies and contributed to the website development for the lab using Next.js.• Designed a system for live transcription and summarization using transformer based models, Achieved an 8% reduction in word error rate through fine-tuning a pre-trained sequence-to-sequence model for multilingual speech recognition.• Achieved near SOTA results in the detection of… Show more • Conducted research on deep learning and transfer learning, designed, and implemented models integrating transfer learning methodologies and contributed to the website development for the lab using Next.js.• Designed a system for live transcription and summarization using transformer based models, Achieved an 8% reduction in word error rate through fine-tuning a pre-trained sequence-to-sequence model for multilingual speech recognition.• Achieved near SOTA results in the detection of some diseases through feature selection using Nature-inspired Algorithms like Grey Wolf Optimization (GWO), implementing advanced ensembling techniques and image segmentation using U-Net architecture with different backbones. Show less

• Developed Jupyter Notebooks to enhance the data science curriculum, explicating diverse ML models, resulting in a 20% improvement in feedback scores compared to previous iterations. Collaborated with industry experts to design masterclasses for each topic, detailing real-world and industry-specific use cases for various models. Engineered an automated scoring system for Python lab test notebooks and ensuring efficient feedback mechanisms.