Tajallah S.

With a strong foundation in neural networks, I've successfully implemented Convolutional Neural Networks (CNNs) and fine-tuned advanced language models (LLMs) like BERT, T5, and GPT for diverse downstream tasks. My expertise spans web development, data engineering, and data science, utilizing tools such as Informatica, Azure Data Factory, Databricks, Spark, Hadoop, PostgreSQL, SQL, MongoDB, and more. In the field of natural language processing (NLP), I've explored sequential models, RNNs, LSTMs, GRUs, and cutting-edge transformers. My academic journey culminated in a thesis focused on machine learning applications in medical data, where I delved into advanced medical image segmentation techniques. This included pioneering work on orthogonal medical image annotations using sparse data through a dual-network paradigm, Dense-Sparse Cotraining (DeSCO). Additionally, I explored federated learning methodologies like FedAvg and FedGen. Proficient in Python, PyTorch, NumPy, TensorFlow, Matplotlib, SQL, Git, Scikit-learn, NLTK, HuggingFace, OpenAI and LangChain, I am well-equipped to contribute to impactful projects at the convergence of web development, data engineering, and machine learning.