Title: Blood Pressure Estimation from Photoplethysmography (PPG) using Deep Neural NetworksIn this thesis work, the heart rate (HR) and blood pressure (BP) was estimated from the raw PPG signals using deep neural networks architecture. • A detailed research work was carried out to analyze the challenges in traditional blood pressure estimation methods. • The genuine patients record datasets was extracted from the MIMIC-III Waveform database using WFDB software package.• The downloaded physiological signals and time series data are pre-processed to remove the artifacts using Python.• The deep neural network framework was developed, which involves convolutional neural network (CNN) and long short-term memory (LSTM) models using PyTorch.• The proposed model was trained and implemented to measure the patient’s heart rate, systolic and diastolic blood pressure from the photoplethysmography (PPG) signals.• The model was evaluated on the unseen test datasets using regression metrics. • The comparison work of small sample and large datasets of patient’s data results are presented and concluded.Software Used: Python, PyTorch, TensorFlow, NumPy, WFDB and HeartPy.

The primary tasks in this position are described below.• Evaluation of welding research data, which are collected experimentally.• Software development in the programming language Python with appropriate tools – Git /GitHub, automated API documentation, CI / CD pipelines. • Creation and updating of software documentation via readthedocs.org.• Creation of tutorials for the developed software using Sphinx.Software Used: Python, GitHub and Sphinx.

The primary tasks in this project were to classify the DNA sequences using deep neural network architecture based on the production turnaround time (manufacturing) of DNA fragments strings. • Understanding and analyzing the data within the company’s database system.• Methods to initialize parameters to build the proposed model.• Build and training the deep convolutional neural network (CNN) to extract high-level features from the raw DNA sequences.• Recognizing the patterns and predicting the sequences based on the label (production time)• Evaluation of the model performance using measuring metrics and testing on the unseen datasets.Software used: Python, Keras & TensorFlow

• Analyzing data such as budgets, and sales results of the company products. • Developing strategies based on the trends in sales data to improve the performance of the organization. • Conducting meetings and presentations to share ideas and performing requirements analysis.

• Developing business strategies and requirements which is a B2B scenario. • Support all functions of sales compensation, including modeling, roll-out process, digital products online campaigns, and sales incentive programs.