Shivachetan Ulavi

adk9go@r.postjobfree.com 781-***-****

https://www.linkedin.com/in/shivachetanulavi/

Master of Science in Business Analytics candidate at Bentley University with an expertise in predictive modelling, handling multiple projects with short deadlines and ability to interpret the results cogently. Passionate about Computer Vision and Artificial Intelligence.

MULTIVARIATE REGRESSION, LOGISTIC REGRESSION, TIME SERIES ANALYSIS & FORECASTING, DECISION TREES, DEEPLEARNING, CLUSTER ANALYSIS, NATURAL LANGUAGE PROCESSING, ANAMOLY DETECTION, CONVOLUTION NEURAL NETWORKS, GRADIENT BOOSTING, DATA WRANGLING, DATA VISUALIZATION TECHNICAL: PYTHON, R, SQL, SAS, OCTAVE, EXCEL & TABLEAU MASTER OF SCIENCE IN BUSINESS ANALYTICS

Bentley University, Waltham, Massachussets Graduated in 2020 3.7 GPA BACHELOR OF ENGINEERING IN COMPUTER SCIENCE

Dr. Ambedkar Institute of Technology, Bangalore, India Graduated in 2017 8.25/10 GPA RESEARCH ASSISTANT

May 2020 - Dec 2020 Mathematical Sciences Department Bentley University, Waltham, MA Extracted ball by ball cricket data for each game across Indian Premier League & Big Bash league, cleansed and added new hand engineered features to the existing data. Reviewed existing literature, formulated new hypothesis & conducted classical statistical analysis by developing manual functions to find 'change points'.

Time series analysis at the league level with help of the estimated 'change points' in order to quantitatively measure the interesting levels & excitement levels, so as to determine which league to be broadcasted on US television.

DATA SCIENCE INTERN

Aug 2020 - Dec 2020 Bentley Digital Marketing, Waltham, MA Scraping general media & social media websites with the help of Python using packages like Requests & BeautifulSoup & filtering out the relevant content & classifying the data. Pre-processing the unstructured raw data into structured text data with the help of Natural Language Processing tools, In order to analyze the sentiments of the associated stakeholders. Conducting statistical analysis & building machine learning models in order to target the right channels & find right keywords to enhance the brand image & engagement to further increase the traction of Bentley University's programs.

HUMANA-MAYS CASE STUDY

Determining the medicare members who were most likely facing transportation issues, the data included member demographics supplemented by public health data consisting of 70000 observations and 750+ features. PROFESSIONAL SUMMARY

SKILLS

EDUCATION

EXPERIENCE

PROJECTS

Worked on classifying the imbalanced problem by building various models including logistic regression, Decision Trees, Neural Networks, Voting classifiers, Random Forests, Ada Boost & Extreme Gradient Boosting. Iteratively reduced the false positive rate of the classifications by tuning various hyper-parameters the ultimate model had an accuracy of over 87%.

Recommended which factors were most likely contributed for the problem & potential solutions to tackle them. ANALYSIS OF INFLUENZA LIKE ILLNESS IN MASSACHUSETTS GRADUATE COURSE: TIME SERIES ANALYSIS Extracted 10 years worth weekly patients data for flu like illness reported by sentinel providers in Massachusetts from CDC.

Built linear, non linear and bagged models for both high-frequency and high-entropy data & low frequency data to forecast influenza patients for the imminent weeks. Measured the accuracy of the models under various performance metrics and concluded the best 3 models which can be subsequently used by concerned parties to prepare for the upcoming demand for resources. WILL THE LOAN BE REPAID? ONLINE COURSE : UDEMY DEEPLEARNING WITH TENSORFLOW Obtained the historical data of LendingClub an US peer-to-peer lending company, which had 396,000 records and 27 associated variables.

Pre-processing the data to fill in missing values & performing exploratory data analysis to find meaningful insights & finally using Keras built a sequential Artificial Neural network. The model had 89% accuracy & for a given person with certain demographics it can predict whether the loan will be repaid or not.

MACHINE LEARNING

https://www.coursera.org/account/accomplishments/certificate/MBLCRBLL7GRB DEEPLEARNING.AI

https://www.coursera.org/account/accomplishments/specialization/certificate/VWTKRVUEMNUP CERTIFICATIONS