As a Business Analytics Student at UTD with a focus in data science I am highly proficient in Machine Learning, Big data Frameworks and Predictive Analytics for business applications.I've worked as a data science intern at a healthcare IT startup where I have developed analytical tools for the Business intelligence team, worked on automating data pipelines, developed a recommendation engine and worked closely with production deployments. I'm proficient in python, SQL, and R and my goal is to use what I have learned to facilitate better decision making in the organization that I am working for.
-
Experience
• Automated the data extraction process from patient reports by implementing an NLP based algorithm which improved the efficiency of lab technicians by 50%
• Designed an efficient database model to store the data extracted from the patient reports which enabled the creation of productive dashboards by the business intelligence team
• Created a recommendation engine which is instrumental in helping medical professionals prescribe correct dosage to patients based on their genetic information and certain demographic factors
• Computerized the data extraction of genetic information by patient ID using natural language processing from PGX reports which resulted in a decrease in time taken to give inputs to the recommendation engine by 60%
• Designed a RESTful API which takes the output of the recommendation engine and recommends the optimal dosages of the prescribed medications
• Successfully automated the process of secondary analysis in Genome sequencing using python scripts by identifying different requirements and obstacles reducing the time to process a batch of samples by 80%
-
Projects
A Study of External Audit (python)
• Predicted whether a firm is fraudulent or not by analyzing a dataset of 777 firms collected from the Auditor office of India with 98% recall • Automated the process to predict the audit risk score based on the risk factors by utilizing a regression model with 87% accuracy
Product Insights using Amazon Reviews (python)
• Examined 15000+ reviews from the amazon website to build 10 a topic model that accurately identifies the content a review into different topics such as heating, performance, storage, etc.
• The classification of each review into a topic provided insights on the customer’s expectations from the product, the data was scraped from amazon using beautifulsoup4 & selenium
Automated Brain Tumor Detection from MRI images (python)
• Automated the process of brain tumor detection from MRI images by employing image processing and machine learning techniques with an accuracy of 82%
• Applied SVM to classify the input as normal or abnormal and further used Multi-SVM to classify various types of tumors in the abnormal images
Manager Recommendations using FIFA 2019 dataset (SAS)
• Utilized forward, backward and stepwise selection with AIC as the criteria to find the best predictive model to predict the market value of players
• Identified the important skills for each position, the trend of the market value with age using statistical models and visualization in the exploratory data analysis