Yashaswini Joshi

Solving complex business problems through the right mix of machine and mind.

I am MS Information graduate student majoring in Data Science and Analytics at the University of Michigan, Ann Arbor. I have strong data science and software engineering experience in developing applications at scale using programming languages such as Python and Javascript. I have worked on applications of Machine Learning and Deep Learning through strong demonstrated projects in the areas of Natural Language Processing, Search and Recommendation, Data Visualization, and Social Network Analysis. Through my demonstrated history of work, I have developed a rich set of transferable skillsets in applications ranging from fields as diverse as healthcare to E-commerce. I also have authored and published papers while working with strong interdisciplinary teams. I enjoy collaborating with teams and developing novel solutions for challenging problems.

Education

MASTER OF SCIENCE - INFORMATION 2019 - 2021 University of Michigan, Ann Arbor, MI, USA Majoring in Data Science and Analytics.

RESEARCH ASSISTANT 2019 - 2020 University of Michigan, Ann Arbor, MI, USA

• Identified problem areas in instructional material through learning analytics using python that improved students’ performance in exams by 85%.

BACHELOR OF ENGINEERING - COMPUTER SCIENCE AND ENGINEERING 2013 - 2017 B V B College of Technology, Hubballi, Karnataka, IND

RESEARCH EXPERIENCE FOR UNDERGRADUATES 2016 - 2017 B V B College of Technology, Hubballi, Karnataka, IND

• Research Publication: Presented a research paper on Mobile Agent based frequent pattern mining of distributed databases in ICICC, 2017 proceedings published by Springer.

Professional Experience

MACHINE LEARNING INTERN 2020 - 2021 Enterprise Minds Inc., San Ramon, CA

• Built a flask application in Docker Container on Google Cloud Platform for extracting the text from invoice images and verify the authenticity of the text.

• DATA SCIENTIST 2020 - 2020 University of Stellenbosch and Wharton School of Business, Ann Arbor, MI

• Built a Bag-of-words model to predict rise and fall in human conflict levels for a given geographic area in defined time window with 85% improvement over baseline.

• Led the Quantitative team, designed and implemented the database system for storing 200 data sets using MySQL. Through automation of ETL process improved the efficiency by 15%.

SOFTWARE ENGINEER 2017 - 2019 Accenture, Bengaluru, KA, IND

• Developed an E2E downloadable monthly customized invoices functionality using agile practices for a fortune 500 telecommunication client which helped decrease operational time and cost by 50%. I was awarded ”Star of the Month”.
• Automated the Change management activity for creating the change requests for project.Conducted impact analysis, assessed change readiness which helped in creating strategies for project progress.

DATA ENGINEER 2016 - 2017 GDV Pvt Ltd, Australia

• Technical associate leading the designing and developing of a web portal for performing statistical analysis of health care data (EHR) of old aged people in order to draw important insights and information using vital health parameters.
• Designed and implemented an intelligent advisory algorithm which advises the individuals on their health and informs the concerned during the emergency situations. The project was awarded the "Best Project" in it's category.

Projects

DRUG RECOMMENDATION USING REVIEW MINING * Built a neural network model that uses the drug descriptions,conditions, past reviews of customers and then recommend the medicines for each Health Condition of a Patient with 89% accuracy.

PREDICTING VALUE OF JEOPARDY QUESTIONS * Predicted value of Jeopardy Questions based on each questions’ text and answer with the accuracy of 78% using Machine Learning and feature extraction.

CRYPTO-CURRENCY TRADING ANALYSIS * Predicting future price movements from historical data to identify price movement patterns through exploratory data analysis(EDA), Statistical analysis and K-means clustering.

CREDIT CARD FRAUD DETECTION * Built a high performance deep learning model to detect credit card frauds with accuracy of 98% using PCA transformation.

OBESITY ANALYSIS * Time series analysis of obesity rates in adults, and the potential associations between obesity and other factors such as socio-economic factors and smoking habits using exploratory data analysis(EDA) and visualized using d3.

PDF SEARCH ENGINE FOR UN AGENCIES AND NGOS * A simple information retrieval based PDF Search Engine with the NDCG score of 0.8 using BM25L model.

SENTIMENT ANALYSIS FOR SELF-DRIVING CARS * Leveraged 20,000 tweets to develop sentiment analysis model to classify them using supervised Machine Learning models, Phrase modeling and Topic Modeling and compared the results through performance matrices.

AIRBNB REVIEW SUMMARIZER * Built a tool that creates visualization of Airbnb reviews per city using Bag of words,Topic Modelling, and Sentiment Analysis.