MAYUKH MAITRA

adg233@r.postjobfree.com LinkedIn GitHub +1-631-***-**** Stony Brook, New York EDUCATION

STONY BROOK UNIVERSITY, NEW YORK, USA, Master of Science in Computer Science, GPA: 3.68 August 2018 - May 2020 Relevant Coursework: Analysis of Algorithms, Natural Language Processing, Data Science, Computer Vision, operating Systems DELHI TECHNOLOGICAL UNIVERSITY, DELHI, INDIA, B.Tech, Software Engineering, Agg: 72.79%August 2011 - June 2015 TECHNICAL SKILLS

● Languages: Python, C, C++, SQL, R, Node Js, Javascript

● Data Science: SQL, Spark, Pandas, R, MongoDB, Oracle MySQL, SQL Server, Teradata, Excel, sklearn

● ETL: Informatica, Teradata, Unix shell scripting

● Machine Learning: Tensorflow, Pytorch, Keras, nltk scikit, OpenCV, Deep learning

● Full Stack: MongoDB, NodeJS, ReactJs, Express

AngularJs

WORK EXPERIENCE

Data Scientist, Senior Associate - Axtria June 2020 - Present

● Built Markov models for performing cost effectiveness analysis and propensity modeling for patient's health states using Python and Excel

● Performed statistical tests like ANOVA, pairwise T tests, etc on patient diagnosis data using R

● Developed predictive machine learning models like XGBoost, SVM, Logistic regression using scikit learn, pandas, etc. to predict a patient's health state by age; Tech stack: Python, SQL, R, Excel, Machine learning, sklearn, matplotlib, statistical modeling Technology Analyst - ZS Associates July 2015 - March 2018

● Performed data engineering on large scale pharma data including Rx Sales, patient data, shipment data, alignment data etc. Designed and developed ETL pipeline for the entire warehouse. Currently hosts data for 4 drugs brands and serves a business of more than $2 million

● Performed logistic regression to estimate the best method for contacting a physician to endorse the drug with the highest conversion rate using Python, sklearn, matplotlib. Improved conversion rate by 67% by optimizing contact strategies

● Implemented complex SQL queries to extract and modify data from various patient data sources and come up with relevant KPIs to create qlikview dashboards which assisted executives of client to track performance of sales

● Undertook and directed the transition process of a periodical reporting project from a different firm. Supervised more than 10 associates for implementation and support of the reporting system. Automated end to end report generation starting from ETL process to PDF report generation and took responsibility for project delivery.

● Tech stack: Python, SQL, Excel, Machine learning, sklearn, matplotlib, Teradata, Informatica, VBA, Unix, Autosys Full Stack Developer - Escale Solutions, India May 2018 - August 2018

● Developed REST APIs for ecommerce websites and product affiliate websites leveraging MEAN Stack, currently catering to more than 1.5 million monthly users and house more than a million products across various categories

● Implemented product recommendation system using content based methodology along with administering customer authentication systems and secure payment gateways

● Spearheaded website and application backend design phase. Designed and implemented database structure utilizing MongoDB. Tech stack: MongoDB, NodeJS, Python, Angular, Javascript, Bootstrap, Express DATA SCIENCE AND MACHINE LEARNING EXPERIENCE

Detecting Rotator Cuff Tears in MRI, Stony Brook University, Advisor: Pravin Pawar March 2019 - June 2019

● Performed local and global feature extraction including HOG, 2D-SIFT and 2D-SURF on rotator cuff MRI images

● Applied various machine learning classification models including SVM, KNN as well as 2D-CNN to classify scans as torn or normal. Achieved best accuracy of 89.14% leveraging HOG on top of Hough transformation feature vector and SVM model Tech stack:OpenCv, KNN, SVM, Deep Learning, Tensorflow, Python, Pandas, scikit-learn; Code Detection of fraudulent credit card transactions (@SBU), Advisor: Steven Skiena March 2020 - May 2020

● Developed a predictive model using xgboost to classify the credit card transaction as fraudulent or non-fraudulent.

● Performed Time-series analysis of sales to find peak transaction time periods and establish correlation using pearson, kendall and spearman with location, holidays and time of day. Visualized the same using matplotlib.

● Reported RMSE of 0.023 and accuracy of 97.68. Tech Stack: Xgboost, Matplotlib, pandas, multinomial logistic regression; Code

Analysis and prediction of real estate price (@SBU), Advisor: Steven Skiena March 2020 - May 2020

● Built a scoring function to rank houses by “desirability”, also designed a Gower distance function to find similar houses using DBSCAN clustering algorithm. Developed single-variable regression models using 10 variables, performed permutation test on each to determine p-value of how good predictions of housing prices were.

● Implemented binary logistic regression based xgboost classifier to predict housing prices. Obtained RMSE of 6.157387983790176e-11; Tech Stack: Xgboost, Matplotlib, pandas, multinomial logistic regression; Code Sentiment Analysis with entity recognition using LSTM and BERT, Advisor: N. Balasubramanian October 2019 - December 2019

● Built an entity based recommendation system by performing named entity recognition and sentiment analysis on Yelp reviews

● Executed feature engineering and training on various Machine and deep learning models including LSTM and BERT and attained best 98.5% using LSTM with glove embedding. Tech stack:BERT, LSTM, Tensorflow, Pytorch, Scikit, Python; Code Content based and collaborative filtering based recommendation system, SBU March 2020 - May 2020

● Implemented content based and collaborative filtering based recommendation system on more than 3.5 million records using pyspark. Executed Jaccard similarity and min hash LSH techniques and achieved 0.95 accuracy. Tech stack:Pyspark Code

Contact this candidate