Not Suppose to disclose publicly

Not supposed to disclose publicly.

Clinical coding is the translation of written, scanned, and/or electronic clinical documentation about patient care into code format. These clinical codes are used by health care providers, government health programs, private health insurance companies, software developers, and others for a variety of applications in medicine.

This coding provides high-quality data for clinicians which ensures better and safer patient care, statistical analysis of diseases and therapeutic actions, to file patient data for insurance purposes and can be used as direct surveillance of epidemic or pandemic outbreaks,

Currently we work on 4 major Taxonomies which are conditions, medications

• Used Binary Classifier Machine learning (ML) model and Multi-Class classifier ML Model to classify clinical text.

• Developed different algorithms to detect specific taxonomy out of clinical test using Elastic search and spacy libraries.

• Developed complete custom tokenizer from spacy default tokenizer.

• Used Rule Based Matching & Entity Ruler to detect sentence formation pattern and to detect unit and values from observations clinical text.

• Developed a custom trained semantic dependency parser model to boost accuracy in detecting required medical parameters from given text.

• Developed this Project from scratch and leading the team to maintain and improve it.

Patient data is the 360 degrees view of the patient which includes Physician Interactions, Prescriptions, Procedures, Discharge, Appointments, Registration, Billing, Readmission, etc., This data leverages advanced analytics and Machine Learning to bring key insights to providers, physicians, and helps them achieve their goal of providing high value, personalized care to their patients.

• Worked in developing a code to populate some important medical parameters and Episodes of patient’s year-wise data into JSON format which helps other analytic engines to get further insights.

• Developed a module which checks Quality check (Both Syntactic & Semantic checks) of generated patient’s data.

• Parallelized the whole project to process more number to patients.

• Developed a module to visualize different aspects from patient’s data using the Bokeh library.

• Received Bright spot award as appreciation from the organization for my work in this project.

This project is to enable an API with custom rules to check if the patient has been suspected of a chronic issue, so it will be useful for the physicians to further treat the patients.

• Worked with the team in developing a custom algorithm that takes patient data as input and gives suspects of chronic diseases (if any) w.r.t rules given by the medical coding team.

• Developed a custom unit testing framework to test changes done before every build.

• Parallelized the process to enrich more number of patients data.

This is an internal project of the HPS Hyderabad wing which is used to generate Pay slips from their invoices & track employee’s attendance and their IN/OUT timings.

• This project was completely developed by me from scratch till deployment in Raspberry Pi and received the Bright spot award as appreciation from the organization.

• Developed modules using SQLite & SQLalchemy python libraries. • Familiar with HTML & CSS to generate our organization’s payslip format.

• Led the team to maintain and improve it further.

To handle the problems caused by occlusions, built a deep learning model to track the people into and out of Queue using Transfer learning techniques in unsupervised way to detect persons from the given feed.

• Involved in building deep neural network using pre-trained model Yolo-V3 of object detection.

• Detected the persons using live camera feed.

• Fed it to model to detect persons only by making necessary changes in model.

• Coded the service and wait time of each person to handle occlusions.

To build a desktop-based application where we can identify a player from a video using Object Detection techniques.

• Collected various images of subject.

• Prepared data by augmenting it to train the model well.

• Built Convolutional neural network model and training it with the subject overall images.

• Testing the built model on video to detect the subject correctly.

COM is the Operation and Maintenance (COM) component in the Component Based Architecture (CBA), which is one implementation of the Ericsson Framework Architecture. COM helps different node and pay loads to be managed by single software. Instead of using different software for configuring different node form different vendor COM is single software for managing the nodes.

• Involved in developing different modules of project.

• Solving Trouble Reports (TR’s) and artifacts.

• Involved in Requirements study and understanding the functionality.

• Involved in preparing the test cases according to the requirements.

• Releasing PRA's after testing the fixes.

• Creating Jenkins jobs for continuous integration.

• Raising bugs in MH Web and providing proper observation and logs to identify & reproduce the bugs.