Seasoned Data Scientist with extensive experience in patient-level data analytics and advanced analytics within the pharmaceutical industry. Skilled in machine learning, statistics, probabilities, problem-solving, data visualization, and programming. Hands-on programming skills specializing in Python, R, SQL, and Matlab.
Industry and Project Experience
⦿ Marketing Al: Built a brand agnostic Deep Learning predictor to identify the next best action (content and channel) for a prescriber based on previous sales, mcm engagement, claims, calls, physician attributes helping the brand and marketing team to have better conversations with the prescribers.
⦿ Built a Cl/CD pipeline for an end-to-end run for Physician Cohort creation, Data Preparation, Data Engineering, Model Data Preprocessing, Model Building using MLflow, distributed hyperparameter tuning using hyperopt and SparkTrials, and application of Business rules to the final recommendations.
IT Consulting Firm, ArtificiaI Intelligence Intern May 2019 – Dec 2019
⦿ Face Detection: Created deep learning models for Face Detection, Face Recognition, and Emotion Detection. Performed face clustering to cluster unidentified faces. Created an end-to-end pipeline on AWS using Lambda functions, SNS, SQS, and each of the above models on EC2 as a service. Obtained an accuracy of 98% for emotion detection, 92% for face detection, and 87% for face recognition.
⦿ Object Detection Engine: To detect different objects (trees, fire hydrants, buildings, cars, etc) of interest in a 360-degree street view image using different CNN architectures and using transfer learning on a pre-trained model to create inference scripts as a service. Accuracies for different objects varied across 70% to 93%
⦿ Crime Classifier: Working on the creation of a crime classifier to detect the type of crime (assault, robbery, violence, etc) from videos. Using ConvLSTMs to detect and classify the type.
Pharma Consulting Firm, Decision Scientist Jun 2016 – Jun 2018
⦿ Healthcare Industry: Worked for one of the largest pharmaceutical companies in the world. Used Real-World Evidence (RWE) to solve a wide miscellany of analyses including Baseline characteristics, patient behavioral patterns, HEOR (Health Economics and Outcomes Research) analysis, and HCRU (Health Care Resource Utilization) analysis using different kinds of data sources like EMR (Electronic Medical Record), Claims data, Survey data, Pragmatic clinical trials to create dashboards in Tableau.
⦿ Retail Industry: Worked with one of the leading retail clients in the world for online customers, to predict their Next Best Action (NBA) using Markov and Bandit Algorithms. Worked on developing more efficient ways to manage stock inventory and help in creating a seamless return program which reduced their restocking costs by 74%.
⦿ Bl Tools: Created multiple fully functional end-to-end dashboards using multiple Bl tools like Tableau, MicroStrategy, SAP Design Studio. Worked with advanced functionalities like integrating with R and Python to perform difficult calculations, creation of advanced visualizations like Sankey charts, Sunburst charts.