• Engaged in the management of clinical data, including data extraction and cleaning for datasets exceeding 1,000 entries. Tasks involved outlier removal and missing value imputation using various techniques such as row deletion, mean substitution, or K-nearest neighbors (KNN).• Developed compelling data visualizations to uncover trends, patterns, and anomalies within intricate datasets.• Implemented advanced machine learning algorithms to forecast future trends and behaviors in data.

Collaborated with cross-functional teams on company-wide projects, playing a pivotal role in mentoring select team members. Seamlessly worked with colleagues to deploy machine learning models effectively.• Non-Invasive Blood Glucose Estimation: Developed a signal processing-based algorithm for healthcare monitoring, marking its first deployment in the country.• Computer-Assisted Sperm Analysis (CASA): Utilized image processing and computer vision algorithms to preprocess sperm images, followed by training deep learning models on the processed images.Subsequently integrated the trained models to quantitatively assess sperm count and classify them into ’normal’ and ’abnormal’ categories.• Sperm Head Detection: Manually curated a dataset using Roboflow and trained a YOLOv5 model onPyTorch. Leveraged transfer learning for fine-tuning, achieving high accuracy in detecting sperm heads.• Daddy Check: Utilized image processing and computer vision techniques for data preparation, followed by training deep learning models to quantify and analyze sperm tracking within the Daddy Check gadget.• ECG Monitoring: Developed a novel approach for visualizing Electrocardiogram (ECG) data using Image Processing algorithms, eliminating the need for specialized medical equipment by utilizing smartphone video recordings.

Designed and delivered an online Natural Language Processing (NLP) course, consisting of 18 recorded sessions.The course garnered significant interest, attracting over 200 students and receiving highly positive feedback. Conducted a series of hands-on projects during the course, including:- Tokenization- Word Embedding- Deep Learning (RNN, LSTM, GRU)- Transformers (BERT)Completed the following projects during the course:1- Conducted a comprehensive comparison and evaluation of four Persian tokenizers (hazm, parsivar, nltk, transformers) for text processing.2- Developed, trained, and rigorously tested word embedding models, evaluating their performance using Analogy tests and various vector similarity metrics.3- Implemented and compared six distinct recursive models for sentiment analysis, both with and without pretrained models.4- Successfully fine-tuned the BERT model for Named Entity Recognition (NER) tasks, achieving state-of-the-art results in entity recognition challenges.